Two days of connection, reflection, and shared ambition. On October 15-16, VMO had the pleasure of hosting ALTEN Asia leaders in Yen Tu, Vietnam, for the ALTEN Asia Seminar 2026 - an opportunity to exchange perspectives, strengthen connections across the region, and explore how we can work more closely as one team. Guided by the theme “Breathe - Rethink - Improve”, the seminar offered time to step back from our daily routines, challenge existing ways of thinking, and discover new perspectives on how we collaborate and move forward together. Beyond the discussions, the peaceful surroundings of Yen Tu provided a memorable setting for meaningful conversations and shared experiences, bringing colleagues closer together beyond the meeting room. A heartfelt thank you to all participants for your openness, engagement, and valuable contributions, and to everyone who helped make this seminar a success. As we continue our Asia First journey, let’s turn shared perspectives into stronger connections, better collaboration, and greater value for our customers. Together, we build tomorrow’s world today.

On October 15-16, VMO had the pleasure of hosting ALTEN Asia leaders in Yen Tu, Vietnam, for the ALTEN Asia Seminar 2026 - an opportunity to exchange perspectives, strengthen connections across the region, and explore how we can work more closely as one team.

ALTEN ASIA SEMINAR 2026 | YEN TU, VIETNAM

Two days of connection, reflection, and shared ambition.

alten-asia-seminar-2026-|-yen-tu-vietnam

As the technology industry continues to evolve, continuous learning and practical skill development are becoming increasingly important for IT engineers. At VMO,&nbsp;IT Education and IT engineer development are an important part of the company’s talent development strategy, with a focus on strengthening technical foundations, practical capabilities, and readiness for international projects. In line with this direction, VMO recently welcomed&nbsp;LPI-Japan at its headquarters for an exchange on IT Education, certification, and approaches to engineer development, exploring potential opportunities for collaboration in talent development.<h2 style="text-align:justify;">1. Proactively developing IT engineers through IT education</h2>VMO is strengthening its focus on&nbsp;IT Education and continuous skill development, combining structured learning with practical experience and on-the-job training (OJT).The company has expanded its education activities through collaboration with Japanese enterprises, universities in Vietnam, PTIT, and other educational partners. By connecting education with real-world experience, VMO aims to provide engineers with continuous opportunities to strengthen their technical knowledge and practical capabilities.<h2 style="text-align:justify;">2. Strengthening engineer capabilities for the Japanese market</h2>A key focus of VMO’s&nbsp;IT engineer development strategy is preparing engineers for the Japanese IT market and offshore projects.The approach emphasizes both&nbsp;fundamental IT skills and practical capabilities, while supporting engineers in learning, developing their skills, and pursuing professional qualifications. VMO also aims to improve the visualization and standardization of engineering skill levels, helping better align capabilities with real-world project requirements.This direction shares common ground with LPI-Japan’s approach, which emphasizes fundamental technology knowledge, practical experience, and step-by-step career development from basic skills to advanced system design and architecture capabilities.<h2 style="text-align:justify;">3. Exchange with LPI-Japan on IT certification and education</h2>VMO recently welcomed&nbsp;LPI-Japan to its headquarters for an exchange on certification systems and IT education frameworks, exploring potential collaboration in engineer development and certification support.Established in 2000 with the participation of major Japanese companies including NEC, Hitachi, Fujitsu, and IBM, LPI-Japan has built a community of more than&nbsp;2,100 engineers, with approximately&nbsp;440,000 examinations conducted and 160,000 certifications issued.During the exchange, LPI-Japan introduced its certification and education approach, while VMO shared its&nbsp;IT Education Promotion Activities with Japanese enterprises, Vietnamese universities, PTIT, and other educational partners.The two sides identified shared perspectives on&nbsp;IT education, engineer development, and human resource development, and will continue exploring potential pilot initiatives and concrete opportunities for collaboration.<h2 style="text-align:justify;">4. Looking Ahead</h2>VMO remains committed to strengthening&nbsp;IT engineer development, building capabilities that meet the evolving needs of the Japanese IT market and global projects. Through continued collaboration with educational and industry partners, VMO will create more opportunities for engineers to learn, grow, and build their careers.

As the technology industry continues to evolve, continuous learning and practical skill development are becoming increasingly important for IT engineers. At VMO, IT Education and IT engineer development are an important part of the company’s talent development strategy, with a focus on strengthening technical foundations, practical capabilities, and readiness for international projects. In line with this direction, VMO recently welcomed LPI-Japan at its headquarters for an exchange on IT Education, certification, and approaches to engineer development, exploring potential opportunities for collaboration in talent development.

VMO advances IT education through exchange with LPI-Japan

vmo-advances-it-education-through-exchange-with-lpi-japan

AI is getting smarter. But without connected data and systems, even the most advanced healthcare AI can struggle to create real-world value.Artificial intelligence is rapidly becoming part of healthcare from medical imaging and predictive analytics to clinical assistants and automated operations.But as healthcare organizations move from AI pilots toward enterprise adoption, a bigger challenge is emerging:Can AI actually work across the fragmented systems where healthcare data lives?The next phase of healthcare AI may depend not only on better models, but on better interoperability.<h2>Healthcare Has Data. The Problem Is Connecting It.</h2>Healthcare organizations generate enormous amounts of information across electronic health records (EHRs), laboratory systems, medical imaging, pharmacy platforms, connected medical devices, wearables, insurance systems, and patient applications.Yet much of this information remains fragmented.Different systems may use different data structures, terminology, identifiers, and integration methods. As a result, an AI solution may have access to only part of the information required to make its output useful.Consider an AI system designed to identify patients at risk of deterioration.To generate meaningful insights, it may need to combine medical history with laboratory results, medications, vital signs, previous admissions, and potentially real-time data from connected devices.If these systems cannot communicate effectively, AI sees only part of the patient picture.<h2>From AI Pilots to Enterprise AI</h2>This problem becomes more visible as healthcare organizations scale AI.Many pilot projects operate in controlled environments with carefully prepared datasets. Enterprise healthcare environments are very different.AI must interact with existing applications, legacy infrastructure, multiple data sources, security policies, and established clinical workflows.That means scaling AI requires more than model performance.It requires an architecture that allows data to move securely and consistently between systems.Standards such as&nbsp;HL7 FHIR are helping healthcare organizations create more standardized ways to exchange clinical information, while standards such as&nbsp;DICOM continue to support medical imaging interoperability.But technical connectivity alone is not enough.Two systems might successfully exchange information while interpreting that information differently.For AI,&nbsp;semantic interoperability, ensuring data has consistent meaning across systems can be just as important as connecting the systems themselves.<h2>AI Agents Raise the Stakes</h2>The rise of AI agents makes interoperability even more important.Traditional AI might analyze information and provide a recommendation. An AI agent could potentially interact with several systems to complete parts of a healthcare workflow.For example, an AI agent supporting patient discharge could retrieve clinical information, review medication and laboratory data, prepare documentation, identify follow-up requirements, and interact with scheduling or communication systems.The model's intelligence is only one part of that process.The agent also needs reliable APIs, consistent data, secure authentication, appropriate permissions, and clear governance over what it can access and what actions it can perform.Without these foundations, adding more AI could simply add another layer of complexity to already fragmented healthcare environments.<h2>Healthcare Needs an Integration-First AI Strategy</h2>For healthcare leaders, the conversation may therefore need to shift from:“Which AI solution should we implement?”to:“Is our technology environment ready for AI to operate across it?”This means looking beyond individual AI applications and strengthening the underlying foundation: data architecture, APIs, interoperability standards, data quality, identity management, cybersecurity, and governance.It also means avoiding another generation of disconnected solutions.If every department deploys its own AI tools, integrations, and data pipelines, healthcare organizations risk creating new AI silos on top of existing technology silos.A more scalable approach is to build reusable data and integration capabilities that can support multiple AI use cases.<h2>The Future Is Connected Intelligence</h2>Healthcare AI will increasingly depend on how effectively intelligence can operate across the wider healthcare ecosystem.Predictive analytics becomes more useful when it can access comprehensive patient information.AI assistants become more valuable when they can securely retrieve information across clinical systems.Remote patient monitoring becomes more powerful when connected-device data can flow into clinical records and trigger meaningful workflows.In this environment,&nbsp;interoperability is no longer simply an IT integration issue. It is becoming a fundamental part of AI strategy.At&nbsp;VMO, an ALTEN company, we support healthcare and technology organizations across the foundations needed for scalable digital solutions, including data engineering, cloud, AI/ML, application development, APIs, and system integration. &nbsp;

Interoperability Is the Next AI Challenge in Healthcare

interoperability-is-the-next-ai-challenge-in-healthcare

Medical imaging is one of the most promising applications of AI in healthcare. From detecting tumors to identifying abnormalities in X-rays, CTs, MRIs, and ultrasound scans, AI can analyze vast amounts of visual data with impressive speed and accuracy.But accuracy alone doesn’t guarantee clinical value. For AI-powered imaging to make a real impact, it must also integrate into clinical workflows, provide explainable results, perform reliably across diverse settings, and ultimately help clinicians make better decisions.<h2>From “Can AI Detect It?” to “Can Clinicians Use It?”</h2>The early conversation around medical imaging AI was largely focused on performance.How accurately can the model identify a tumor? How well can it distinguish between benign and malignant findings? Can it detect abnormalities earlier than a human reader?These questions remain important. But they represent only one part of the challenge.Imagine an AI system that identifies a potential abnormality with 95% accuracy. If the result appears as another notification in an already overloaded workflow, requires clinicians to open a separate platform, or provides no useful context about why the finding matters, its clinical impact may be limited.The real value of medical imaging AI is not simply detecting something. It is helping clinicians understand what they are seeing and decide what to do next.That requires AI to move beyond isolated image analysis and become part of the broader clinical decision-making process.<h2>The Clinical Workflow Matters</h2>Healthcare professionals rarely look at medical images in isolation.A radiologist may consider a patient's previous scans, clinical history, laboratory results, symptoms, medications, and other information before concluding. An imaging AI system that only analyzes a single image sees just one piece of the puzzle.This creates an important opportunity for AI.Instead of simply highlighting an area of concern, future systems can potentially combine imaging data with relevant clinical information to provide a more comprehensive picture.For example, an AI system could flag a suspicious lung nodule while also comparing it with previous scans, measuring how it has changed over time, and surfacing relevant patient information.<h2>Explainability Is Becoming Essential</h2>Trust is another critical challenge. When AI influences clinical decisions, healthcare professionals need more than a prediction they need to understand why the system reached its conclusion.Explainable AI can help by highlighting abnormalities, showing confidence levels, and providing context behind its recommendations. The goal is not to replace clinical judgment, but to give clinicians the insight they need to&nbsp;evaluate, question, and confidently use AI recommendations.<h2>Accuracy Can Change Across Real-World Settings</h2>A model that performs exceptionally well on one dataset may not perform the same way in another hospital.Different hospitals may use different imaging equipment, protocols, patient populations, image quality standards, and workflows. Even relatively small differences in data can affect how an AI model performs.This is why&nbsp;generalization matters as much as benchmark accuracy.Healthcare organizations need to ask:<ul><li>Does the model perform consistently across different devices and hospitals?</li><li>Has it been validated on diverse patient populations?</li><li>How does performance change when image quality is poor?</li><li>How frequently does the system need to be monitored or recalibrated?</li></ul>These questions become especially important as healthcare AI moves from pilot projects into enterprise environments.An impressive accuracy score is useful. But&nbsp;consistent performance in the real world is what creates clinical confidence.<h2>The Human Factor Still Matters</h2>Better AI doesn’t mean less need for human expertise. Instead, AI and clinicians can complement each other.AI can process large volumes of images, identify patterns, prioritize cases, and reduce repetitive tasks, while clinicians provide context, experience, and patient-centered judgment.As imaging volumes grow, this collaboration can help clinicians focus more on complex cases and better decision-making.<h2>From Diagnostic Tool to Intelligent Imaging Platform</h2>The next stage of medical imaging AI may therefore be less about building a single highly accurate model and more about creating an intelligent imaging ecosystem.Such a platform could connect imaging data with electronic health records, historical examinations, clinical workflows, and other sources of patient information.Instead of AI operating as a standalone tool, it becomes part of the infrastructure through which clinicians access and interpret medical information.This also changes what healthcare organizations should evaluate when investing in AI.Accuracy should remain a core metric, but it should sit alongside factors such as interoperability, workflow integration, explainability, scalability, security, governance, and ongoing model monitoring.<h2>What Comes Next?</h2>A model that achieves impressive results in a laboratory is only the starting point. Real impact comes when that technology works reliably across different environments, integrates into clinical workflows, earns the trust of healthcare professionals, and helps turn complex data into better decisions.In other words,&nbsp;the future of medical imaging AI will not be defined by accuracy alone.&nbsp;It will be defined by how effectively intelligence is translated into clinical value.For healthcare organizations exploring AI adoption, this is an important shift in mindset:&nbsp;don't just ask how accurate the model is. Ask what happens after the model makes its prediction.That is where the real value of AI begins.

AI-Powered Medical Imaging: Why Accuracy Alone Isn’t Enough

ai-powered-medical-imaging-why-accuracy-alone-isn’t-enough

Healthcare AI is moving beyond the chatbot.For years, conversational AI has helped patients find information, schedule appointments, and communicate with healthcare providers. But the next wave of healthcare transformation is focused on something much more powerful:&nbsp;AI agents that can understand context, make decisions within defined boundaries, and take action across clinical workflows.Hospital AI agents are emerging as a new layer of intelligent automation—helping healthcare organizations reduce administrative burden, accelerate clinical operations, and give healthcare professionals more time to focus on patient care.<h2>From Chatbots to AI Agents</h2>Traditional healthcare chatbots are primarily designed to respond to questions.A patient asks, “When is my appointment?” and the chatbot retrieves the relevant information. A patient asks about hospital services, and the chatbot provides an answer. AI agents operate differently.Rather than simply responding, an AI agent can&nbsp;observe a situation, reason through a workflow, interact with multiple systems, and execute predefined actions.For example, when a patient needs a follow-up appointment, an AI agent could:<ol><li>Review the patient's relevant clinical information.</li><li>Identify the appropriate follow-up protocol.</li><li>Check physician and department availability.</li><li>Contact the patient through their preferred channel.</li><li>Schedule the appointment.</li><li>Update the hospital's systems.</li><li>Notify the care team when necessary.</li></ol><h2>Where Hospital AI Agents Can Make an Impact</h2>AI agents can support a wide range of hospital operations, particularly workflows that involve repetitive tasks, multiple systems, and large volumes of information.<h3>1. Clinical Documentation</h3>Documentation remains one of the most time-consuming activities for healthcare professionals.AI agents can assist with collecting information from clinical conversations, structuring notes, summarizing patient histories, and preparing documentation for physician review.Instead of replacing clinical judgment, the agent acts as an intelligent assistant that helps reduce the administrative workload.<h3>2. Patient Flow and Care Coordination</h3>Hospitals must continuously coordinate patients, clinicians, beds, diagnostic services, and other resources.AI agents can monitor patient-flow information and identify potential bottlenecks.For example, an agent could detect that a patient's diagnostic results are available, check the next step in the care pathway, notify the appropriate team, and trigger the required workflow. This creates a more connected process between departments.<h3>3. Scheduling and Follow-Up</h3>Scheduling often involves multiple systems and human coordination.AI agents can automate routine activities such as:<ul><li>Appointment scheduling and rescheduling</li><li>Follow-up reminders</li><li>Pre-visit instructions</li><li>Referral coordination</li><li>No-show follow-ups</li><li>Post-discharge communications</li></ul><h3>4. Clinical Decision Support</h3>AI agents can also support clinicians by bringing together relevant information from different sources.An agent could help summarize a patient's history, medications, laboratory results, imaging reports, and previous encounters before a clinical decision is made.Importantly, these systems should operate within clearly defined clinical and governance boundaries.The goal is not to let AI independently make high-risk medical decisions. Instead, AI can help clinicians&nbsp;access the right information at the right time and reduce cognitive and administrative workload.<h3>5. Revenue Cycle and Administrative Operations</h3>The opportunity extends beyond direct clinical workflows.AI agents can support:<ul><li>Insurance verification</li><li>Claims processing</li><li>Prior authorization workflows</li><li>Medical coding assistance</li><li>Documentation checks</li><li>Billing inquiries</li><li>Patient financial communications</li></ul><h2>The Real Opportunity: Connecting the Hospital</h2>One of the biggest advantages of AI agents is their ability to operate across workflows rather than within a single application.A modern hospital may have an EHR, laboratory information system, radiology platform, pharmacy system, scheduling platform, CRM, patient portal, and numerous other applications.The challenge is that these systems do not always communicate seamlessly. AI agents can act as an intelligent orchestration layer across these environments.For example:Patient request → AI agent → EHR → scheduling system → notification → care teamInstead of requiring staff to manually move information between systems, an AI agent can coordinate the workflow according to predefined rules and permissions.This is where AI agents can move healthcare AI from&nbsp;information retrieval to workflow execution.<h2>AI Agents Need More Than a Large Language Model</h2>Building a hospital AI agent is not simply a matter of connecting a chatbot to an LLM.Healthcare environments require a much broader technology foundation.A production-ready AI agent may need:AI &amp; Machine Learning Large language models, retrieval-augmented generation, clinical NLP, reasoning models, and specialized AI models.Healthcare Integration Integration with EHRs and other clinical systems using standards such as HL7 and FHIR.Workflow Orchestration The ability to coordinate multiple steps, systems, users, and business rules.Security &amp; Identity Role-based access, authentication, authorization, audit trails, and secure handling of sensitive healthcare information.Human-in-the-Loop Controls Critical actions should be reviewed or approved by authorized healthcare professionals where appropriate.Monitoring &amp; Governance Organizations need to monitor agent performance, accuracy, actions, exceptions, and potential risks.The technology architecture must therefore be designed around&nbsp;trust, interoperability, security, and accountability.<h2>Human-in-the-Loop Is Essential</h2>Healthcare is different from many other industries because AI errors can have serious consequences. Hospital AI agents should therefore not be designed around unrestricted autonomy.A better approach is&nbsp;controlled autonomy. AI can independently handle low-risk, repetitive activities while escalating complex or high-risk situations to qualified professionals.For example:Low risk: Confirm an appointment → AI handles it.Moderate risk: Identify missing information in a referral → AI flags it for staff review.High risk: Recommend a significant change to a patient's treatment → AI supports the clinician, but the final decision remains with an authorized professional.This approach allows hospitals to capture the efficiency benefits of AI while maintaining appropriate clinical oversight.<h2>What Hospital AI Agents Could Look Like in Practice</h2>Imagine a patient being discharged from a hospital.Traditionally, several teams may need to coordinate follow-up appointments, prescriptions, documentation, patient instructions, and communication.An AI agent could coordinate much of this workflow.After discharge, it could verify that the required documentation is complete, identify follow-up requirements, coordinate appointments, send personalized instructions to the patient, monitor whether required actions have been completed, and escalate exceptions to hospital staff.<h2>Challenges Healthcare Organizations Must Address</h2><h3>Data Privacy</h3>Healthcare data is highly sensitive. Organizations need robust security, access controls, data governance, and privacy mechanisms throughout the AI lifecycle.<h3>Accuracy and Hallucination</h3>AI-generated information must be validated, particularly when used in clinical environments. Retrieval mechanisms, validation layers, confidence thresholds, and human review can help reduce risk.<h3>Integration Complexity</h3>AI agents are only as useful as their ability to interact with existing healthcare systems. Legacy applications and fragmented data environments can make integration challenging.<h3>Regulatory Compliance</h3>Healthcare AI must operate within applicable regulatory and organizational requirements. Governance should be designed into the architecture rather than added after deployment.<h3>Change Management</h3>Introducing AI changes how healthcare professionals work. Successful implementations require training, workflow redesign, clear accountability, and user trust—not just new technology.<h2>The Future: Agentic Healthcare</h2>The next stage of healthcare AI will not simply be about asking an AI assistant questions.It will be about&nbsp;delegating well-defined workflows to intelligent systems.A patient may interact with one AI agent. A nurse may rely on another to coordinate tasks. A hospital operations team may use agents to optimize patient flow. Administrative teams may use agents to automate financial and operational processes.Behind the scenes, these agents could work together across hospital systems while humans remain responsible for decisions that require clinical expertise and judgment.This represents a shift from:Chat → Assistance → Automation → OrchestrationThe hospitals that successfully adopt this model will not necessarily be those with the most AI tools. They will be those that identify the right workflows, establish strong governance, integrate AI with existing systems, and create a clear balance between&nbsp;automation and human expertise.<h2>How VMO Can Help Healthcare Organizations Build AI-Driven Operations</h2>At VMO, we see AI agents as part of a broader healthcare technology transformation.Building effective AI solutions requires more than AI models. It requires expertise across&nbsp;AI engineering, data engineering, cloud, software development, healthcare integration, cybersecurity, and digital transformation.From intelligent workflow automation and clinical data platforms to AI-powered applications and system integration, VMO can help healthcare organizations explore, build, and scale AI solutions around real operational needs.

Healthcare AI is moving beyond the chatbot.
For years, conversational AI has helped patients find information, schedule appointments, and communicate with healthcare providers. But the next wave of healthcare transformation is focused on something much more powerful: AI agents that can understand context, make decisions within defined boundaries, and take action across clinical workflows.

Hospital AI Agents: Automating Clinical Operations Beyond Chatbots

hospital-ai-agents-automating-clinical-operations-beyond-chatbots

Drug discovery has always been a race against time.From identifying promising biological targets to designing molecules, validating candidates, and moving them toward clinical development, bringing a new drug to market can take more than a decade and require enormous investment. Even after years of research, many candidates fail because they are ineffective, unsafe, or simply impossible to develop at scale.Generative AI is beginning to change that equation.Unlike traditional AI systems that primarily analyze existing data, generative AI can&nbsp;create new molecules, predict their properties, generate hypotheses, and explore potential drug candidates that researchers may never have considered manually.The result is a new model for pharmaceutical R&amp;D: instead of searching through a vast chemical universe one possibility at a time, researchers can increasingly use AI to generate and prioritize the most promising paths.<h2>Why Drug Discovery Takes So Long</h2>Traditional drug discovery involves multiple stages, each with its own bottlenecks.Researchers first identify a biological target associated with a disease. They then search for compounds that can interact with that target, optimize those compounds for potency and selectivity, evaluate their safety and pharmacokinetic properties, and eventually move successful candidates into preclinical and clinical testing.A simplified process looks like this:Disease research → Target identification → Hit discovery → Lead optimization → Preclinical testing → Clinical trials → Regulatory approvalThe challenge is that each stage produces uncertainty.A molecule may bind effectively to a target but have poor absorption. Another may show strong activity in laboratory experiments but prove toxic in animals. A promising candidate may also be too difficult or expensive to manufacture.<h2>From Finding Molecules to Designing Them</h2>One of the biggest changes introduced by generative AI is the shift from&nbsp;molecule screening to molecule generation.Traditional virtual screening evaluates large libraries of existing compounds to identify molecules that could potentially interact with a biological target.Generative models approach the problem differently.They can learn patterns from existing chemical and biological data and then generate entirely new molecular structures designed around specific objectives.For example, researchers could ask an AI system to generate molecules that:<ul><li>Bind strongly to a specific protein target</li><li>Have improved selectivity</li><li>Reduce predicted toxicity</li><li>Have better solubility</li><li>Meet specific pharmacokinetic requirements</li><li>Remain synthetically feasible</li></ul>Instead of asking,&nbsp;“Which molecule in this library might work?”, researchers can increasingly ask:“What molecule should we design for this biological problem?”That is a fundamental shift in the drug discovery workflow.<h2>How Generative AI Is Accelerating R&amp;D</h2><h3>1. Faster Target and Disease Research</h3>Generative AI can process enormous amounts of scientific information, including research papers, genomic data, protein information, clinical data, and proprietary pharmaceutical datasets.This can help researchers identify relationships that might otherwise take significant time to uncover.AI can help answer questions such as:<ul><li>Which biological pathways are associated with a disease?</li><li>Which proteins could represent promising therapeutic targets?</li><li>What evidence supports a particular target?</li><li>Which mechanisms have already been explored?</li><li>Where are the gaps in existing research?</li></ul>Rather than replacing scientists, these systems can act as research copilots that help teams navigate increasingly complex scientific knowledge.<h3>2. De Novo Molecule Generation</h3>Perhaps the most transformative application is&nbsp;de novo drug design.Generative models can create new molecular structures rather than simply selecting from existing chemical libraries.Researchers can then computationally evaluate these candidates before deciding which ones deserve laboratory testing.This can dramatically reduce the number of molecules that need to be synthesized and tested physically.The workflow becomes:Generate → Predict → Rank → Synthesize → Test → RefineInstead of synthesizing thousands of possibilities and discovering which ones work, researchers can potentially focus laboratory resources on a much smaller set of AI-prioritized candidates.<h3>3. Protein Structure Prediction</h3>Understanding the three-dimensional structure of proteins is critical to modern drug discovery.Knowing how a protein folds can help researchers understand where and how a drug molecule might interact with it.Advances in AI-based protein structure prediction have significantly expanded researchers' ability to study biological targets whose structures were previously difficult or expensive to characterize.This creates an important connection between&nbsp;protein intelligence and generative drug design.AI can help researchers understand the target and then generate molecules designed to interact with it.<h3>4. Multi-Parameter Optimization</h3>Drug candidates rarely need to be optimized for only one characteristic.A successful molecule may need to be potent, selective, stable, soluble, safe, orally bioavailable, and manufacturable at a reasonable cost.Optimizing all these properties simultaneously is extremely difficult.Generative AI can help explore this multi-dimensional design space by generating candidates against multiple constraints.This allows drug developers to consider&nbsp;developability earlier, rather than discovering late in the process that a promising molecule has an unacceptable weakness.<h3>5. Smarter Preclinical Decisions</h3>AI can also help researchers predict how candidate molecules may behave before extensive laboratory testing.These predictions are not substitutes for experimental validation. However, they can help researchers identify weaker candidates earlier and prioritize experiments with greater potential value.That can reduce wasted laboratory time, materials, and resources.<h2>The Rise of AI-Native Drug Discovery</h2>Generative AI is also creating a new category of pharmaceutical companies:&nbsp;AI-native drug discovery companies.Rather than using AI as a supporting tool inside a traditional R&amp;D organization, these companies are building their discovery processes around computational models from the beginning.Their approach typically combines:Generative AI + Biological Data + Automation + High-Throughput ExperimentationThe goal is not simply to build better prediction models. It is to create a continuous feedback loop between computation and experimentation.<h2>Why Speed Matters So Much</h2>Reducing drug discovery timelines is not simply about getting products to market faster.Every year removed from R&amp;D can have significant economic value.Pharmaceutical companies invest heavily in research programs where the probability of failure is high. If AI can help identify ineffective candidates earlier, prioritize stronger candidates, and reduce unnecessary experiments, the impact can extend beyond development speed.It could potentially mean:Lower R&amp;D costs + Faster experimentation + Better candidate selection + More efficient use of scientific talentFor patients, the ultimate benefit could be even more important: faster access to new therapies for diseases where current treatment options remain limited.<h2>But Generative AI Is Not a Magic Shortcut</h2>Despite its potential, AI does not eliminate the fundamental challenges of drug development. A molecule that looks excellent computationally can still fail in the laboratory. A successful experiment in a cell model does not guarantee success in animals.The biggest misconception is that generative AI can simply “generate a drug.”In reality, AI can accelerate&nbsp;parts of the discovery and optimization process, but the path from an AI-generated molecule to an approved medicine still requires rigorous scientific validation.<h2>The Future: AI and Scientists Working as One System</h2>The future of drug discovery is unlikely to be humans versus AI. It will be humans&nbsp;with AI.Scientists bring domain expertise, biological intuition, experimental judgment, and the ability to interpret unexpected results. AI brings scale, computational power, pattern recognition, and the ability to explore enormous design spaces.Together, they can create a much faster R&amp;D feedback loop.The pharmaceutical organization of the future may look less like a linear pipeline and more like an interconnected system:Scientific knowledge → AI hypothesis → Generative design → Simulation → Automated experiment → Experimental data → AI learning → New hypothesisThis creates the possibility of continuously improving discovery cycles rather than moving through isolated research stages.<h2>What This Means for Pharma Companies</h2>For pharmaceutical leaders, the question is no longer simply whether generative AI will influence drug discovery.The more strategic question is&nbsp;where AI can create measurable value across the R&amp;D lifecycle.The most successful implementations will likely combine AI capabilities with strong data foundations, domain expertise, cloud infrastructure, and secure engineering environments.<h2>The New Drug Discovery Race</h2>Generative AI is changing the competitive landscape of pharmaceutical R&amp;D.The companies that gain an advantage may not simply be those with the largest research budgets or the biggest compound libraries. Increasingly, the advantage could belong to organizations that can&nbsp;turn data into hypotheses, hypotheses into experiments, and experiments into validated candidates faster than their competitors.And as AI models become more capable, laboratory automation becomes more connected, and biological datasets continue to expand, the traditional decade-plus drug discovery journey may increasingly become a race measured not only in years, but in how intelligently every experiment is chosen.

The New Drug Discovery Race: How Generative AI Is Cutting Years Off R&D
Drug discovery has always been a race against time.
From identifying promising biological targets to designing molecules, validating candidates, and moving them toward clinical development, bringing a new drug to market can take more than a decade and require enormous investment. Even after years of research, many candidates fail because they are ineffective, unsafe, or simply impossible to develop at scale.

Generative AI: Speeding Up Drug Discovery R&D

The New Drug Discovery Race: How Generative AI Is Cutting Years Off R&D

generative-ai-speeding-up-drug-discovery-r-d

Analyze business/technical requirements and mapping documentation.Design and execute test cases, test scripts, and test data for data pipelines.Conduct System Integration Testing (SIT) and support User Acceptance Testing (UAT).Reconcile data between source and target systems across various data pipeline stages.

Automation/Data Test Engineer_Onsite_Hai Ba Trung

automation-data-test-engineer_onsite_hai-ba-trung

1. AI Agent &amp; Knowledge Base Development Build structured, traceable domain knowledge bases from interview notes, qualitative data, and other research sources. Develop Research and Quality Control Agents that can query the knowledge base and provide evidence-backed, auditable results. Design AI workflows for automated QC, import-data classification, natural-language data access, and other repetitive data-processing tasks. Implement robust agent architectures with tool/function calling, RAG, structured outputs, and human-review workflows. 2. AI Platform Integration &amp; Deployment Integrate AI agents with platforms such as ChatGPT, Claude Code, and Google ADK. Configure MCP connections and develop reliable agents capable of completing business tasks. Deploy AI services and integrate them with structured data platforms, including BigQuery. Design and implement end-to-end solutions, including backend services, AI orchestration, APIs, and frontend integration. 3. Evaluation &amp; Productionization Define evaluation criteria and test AI agents against real business requirements. Implement logging, monitoring, guardrails, error handling, and quality-control mechanisms. Improve agent reliability, accuracy, and traceability through iterative testing and optimization. Support the transition from initial pilot to production-ready solutions.

AI Engineer_Onsite_Banking domain

ai-engineer_onsite_banking-domain

Role OverviewWe are looking for hands-on GenAI Fullstack Engineers to design, prototype, and deploy AI-driven solutions that improve internal business processes. This is a small, implementation-focused team. Engineers are expected to work directly on the solution and be able to independently drive a project with a Technical Leader and/or 1–2 other team members.Key Responsibilities<ul><li>Implement GenAI applications hands-on from prototype to production.</li><li>Develop AI pipelines including LLM integration, prompt orchestration, RAG, document processing, APIs, and business-system integration.</li><li>Build backend services, connectors, automation workflows, and supporting components using Python or TypeScript.</li><li>Troubleshoot and optimize AI pipelines for accuracy, latency, reliability, and cost.</li><li>Work directly with the Product Owner to understand requirements and translate them into working solutions.</li><li>Take technical ownership of a specific component or pipeline and be able to explain its implementation in detail.</li><li>Use AI-assisted development tools to accelerate implementation while maintaining code quality and engineering standards.</li></ul>

Fullstack AI Engineer_AI Department

fullstack-ai-engineer_ai-department

• Collaborate with Agile product teams (mobile, digital banking, corporate banking, and customer insight) to design, develop, test, implement, and support full-stack data solutions that directly impact millions of mobile banking users • Design and evolve our enterprise Data Platform with Data Vault 2.0 as the foundational modeling methodology while delivering high-quality, trusted data layers in Google Cloud • Build high-throughput, low-latency real-time and batch data pipelines to power behavioral analytics, next-best-action recommendations, and seamless customer experiences • Create scalable ingestion, transformation, quality enforcement, and serving layers that feed our Next-Gen Mobile Banking App and other Digital Banking platform products • Ensure enterprise-grade data quality, governance, lineage, security, and compliance in a highly regulated environment • Optimize pipelines for performance, cost-efficiency, reliability, and scalability on Google Cloud Platform (GCP) • Mentor junior and mid-level data engineers, conduct design reviews, and leverage DataOps best practices (CI/CD, automated testing, observability) • Partner closely with mobile app teams, data scientists, and business stakeholders to translate digital banking requirements into production-grade data products

Data Engineer_Onsite_Hai Ba Trung

data-engineer_onsite_hai-ba-trung

<ul><li>Software Development: Designing, developing, testing, deploying and maintaining modular- based systems in accordance with best practices. Problem Solving: Troubleshoot incidents, identify root cause, fix and document problems, and implement preventive measures.</li><li>Agile Scrum Practices: Participate actively in Agile Scrum processes, including sprint planning, daily stand-ups, retrospectives, and demos.</li><li>Collaboration: Work closely with cross-functional teams, including product management, design, and QA to ensure seamless integration and delivery.</li><li>Documentation: Maintain accurate and up-to-date technical documentation of software components and systems.</li></ul>

Techlead (C#/.Net) (Onsite Japan)

techlead-c-net-onsite-japan

<ul><li>Correct and convert SQL queries as needed to meet project requirements.</li><li>Gather and analyze relevant information regarding reported issues.</li><li>Trace and identify the origin of specific issues within our data pipelines.</li><li>Provide initial feedback to the issue reporter, ensuring clear and timely communication.</li><li>Document all aspects of the issue, including actions taken, timelines, time spent, and any impediments encountered.</li><li>Summarize findings and escalate issues to the core Data Engineering (DE) team for further action when necessary.</li><li>Follow existing Standard Operating Procedures (SOPs) or develop new ones to address unique challenges.</li><li>Make intuitive connections between data inputs and outputs to identify and resolve potential issues. Bonus points</li><li>Understanding of Operational Technology (OT) domains, such as DCS / APC / soft sensor technologies from Yokogawa, Honeywell, Emerson, ABB, Siemens, AspenTech, or AVEVA, or hands-on experience working in a manufacturing plant.</li><li>Japanese N2+ proficiency or direct experience working with Japanese clients, including Japanese-language review documents, internal notes, and client communications / IMs.</li><li>Understanding of AI-related regulations and standards such as FDA 21 CFR Part 11, EU AI Act, and IEC 62443.</li><li>Experience building REST APIs and web dashboards for data/model visualization.</li><li>Knowledge of reinforcement learning, time-series forecasting, digital twins, or process simulators.</li><li>Exposure to pre-sales &amp; solutioning, or experience interviewing and mentoring other engineers</li><li>For senior candidates: experience designing architecture and leading on engagements &nbsp;</li></ul>

AI Consultant (Onsite Japan)

ai-consultant-onsite-japan

<ul><li>Analyze requirements and work with Developers and the Product Owner to clarify business logic and define test scope.</li><li>Design test scenarios and test cases; run functional, API, integration, and regression tests.</li><li>Build, maintain, and improve automation tests in C#/.NET and NUnit, using the project's shared libraries.</li><li>Test login and access control flows with Keycloak/OAuth, and email flows via SMTP.</li><li>Write, run, and analyze performance tests with k6/JavaScript.</li><li>Prepare and manage test data; query and verify data in SQL Server.</li><li>Run automation tests on Azure DevOps; use YAML/PowerShell to support pipelines and test tasks.</li><li>Test on Azure/Kubernetes environments; read application logs (including NLog) to investigate bugs.</li><li>Log, track, and verify bugs; report test results and keep QA documents up to date.</li></ul>Extra responsibilities for Senior<ul><li>Propose test and automation strategies based on scope, risk, and priority.</li><li>Improve the automation framework, test stability, and CI/CD run efficiency.</li><li>Review test code, mentor team members, and discuss technical issues with the team or client.</li></ul>

Automation Test - Du33 - english - PreSenior / Senior

automation-test-du33-english-presenior-senior

<ul><li>Develop and deploy NLP solutions for Japanese and English text processing.</li><li>Implement and optimize computer vision models for object detection, image classification, and OCR.</li><li>Fine-tune, optimize, and deploy Large Language Models (LLMs).</li><li>Design and implement end-to-end ML pipelines for training, evaluation, and deployment.</li><li>Optimize model performance using techniques such as quantization and pruning.</li><li>Collaborate with research teams to develop and implement state-of-the-art ML solutions.</li></ul><h2>&nbsp;</h2>

Computer Vision (Osaka)

computer-vision-osaka

About the Role We are looking for a Chinese-speaking Comtor who is interested in developing a long-term career as a Business Analyst (BA). The primary responsibility of this role is to support communication between Chinese-speaking clients/business users and the internal project team, ensuring that business requirements, questions, and project information are clearly understood and communicated. For candidates with the right capability, we also expect the opportunity to gradually take on Junior BA responsibilities from day one, including requirement clarification, documentation, process analysis, UAT support, and coordination with Development/QA teams. This is a good opportunity for someone who has a strong interest in IT Software and Business Analysis, is proactive in learning, and wants to grow from a Comtor into a BA.Key Responsibilities 1. Comtor Responsibilities<ul><li>Act as a communication bridge between Chinese-speaking clients/business users and the internal project team.</li><li>Translate and interpret business discussions, requirements, questions, and project-related information between Chinese and Vietnamese/English as required.</li><li>Join meetings with clients and support communication between clients and the project team.</li><li>Ensure key information and requirements are clearly understood by both business and technical teams.</li><li>Support communication and follow-up on project issues, changes, and action items.</li><li>Prepare and maintain translated project-related documents, messages, and communications.</li></ul>2. BA Responsibilities<ul><li>Depending on capability and readiness, gradually take ownership of BA activities such as: Gather, clarify, and analyze business requirements with clients/business users.</li><li>Proactively identify unclear requirements, missing information, assumptions, and potential edge cases.</li><li>Translate business needs into clear and structured requirements for the Development/QA team.</li><li>Prepare BA documentation such as:<ul><li>Business/functional requirements</li><li>User stories</li><li>Process flows</li></ul></li><li>Requirement clarification documents</li><li>Meeting minutes and action items</li><li>Work closely with the Product Owner, Development, QA, and other project members to ensure requirements are correctly understood and implemented.</li><li>Support requirement workshops and discussions with clients.</li><li>Follow up on open questions, requirements, issues, and agreed actions.</li><li>Support SIT/UAT activities, including test scenario preparation, issue clarification, and coordination with relevant teams.</li><li>Understand the impact of requirement changes and communicate them to relevant stakeholders.</li><li>Proactively monitor project progress and raise potential risks or issues when identified.</li><li>Continuously improve understanding of the client's business processes and the software solution. &nbsp;</li></ul>Career Development<ul><li>This role is designed for candidates who want to build a career beyond translation.</li><li>Career path: Comtor → Comtor + Junior BA → Business Analyst → Senior Business Analyst</li><li>As your BA capability develops, you will gradually take more ownership of requirement analysis, client communication, documentation, and project delivery activities. &nbsp;</li></ul>What You Will Gain<ul><li>Hands-on experience working with international clients and Chinese-speaking business users.</li><li>Exposure to real-world software development and IT outsourcing processes.</li><li>Opportunity to work closely with Product Owners, BA, Development, and QA teams.</li><li>Practical experience in requirement analysis, UAT, process analysis, and stakeholder management.</li><li>A clear opportunity to transition from Comtor to Business Analyst based on your capability and performance.</li><li>Exposure to software solutions and business processes in the Pharmaceutical industry.</li></ul>

Chinese IT Comtor

chinese-it-comtor

1. About the Role- We are looking for an experienced Data Scientist, AI Engineer to join our AI team and work on end-to-end AI solutions, from Machine Learning and predictive analytics to Generative AI, NLP, RAG, and Agentic AI.- You will work closely with technical teams and business stakeholders to transform real-world business problems into scalable, production-ready AI solutions.2. Key Responsibilities- Design, develop, and deploy end-to-end Machine Learning and AI solutions, including traditional ML, predictive analytics, time series forecasting, deep learning, and Generative AI.- Analyze complex business problems and translate business requirements into scalable AI/Data Science solutions.- Develop and optimize NLP, LLM, and Deep Learning models for real-world business applications.- Build LLM-powered applications, RAG systems, and Agentic AI workflows.- Develop and maintain production-grade ML/AI pipelines using Python and PySpark.- Conduct data analysis, feature engineering, model development, evaluation, optimization, and deployment.- Collaborate with business stakeholders and technical teams to deliver practical and measurable business impact.- Leverage Generative AI tools and Agentic AI frameworks to improve productivity and automate workflows.- Communicate analytical findings, technical concepts, and AI solutions effectively to both technical and non-technical audiences.- Continuously research and apply emerging technologies and methodologies in AI, Machine Learning, and Generative AI.- Working arrangement: Full-time remote or hybrid.

Automation/Data Test Engineer_Onsite_Hai Ba Trung

Automation/Data Test Engineer_Onsite_Hai Ba Trung

INFORMATION