Principal AI Architect | Principal Data Scientist

Location: Ho Chi Minh City, On-site.
Level: Expert, individual contributor (no people management)
ABOUT THE OPPORTUNITY
ACB is scaling AI and advanced analytics as core capabilities across its business and operations. We are hiring two senior experts who will set technical direction in their field, own high-impact outcomes end-to-end, and act as the bank's reference authority for AI and data science.
====================================================================
POSITION 1 - PRINCIPAL AI ARCHITECT
====================================================================
KEY RESPONSIBILITIES:
Architecture and Strategy:
Define the target architecture, technology roadmap and reference patterns for GenAI, RAG, Agentic and Multi-Agent systems across the bank.
Design and govern the AI landing zone on AWS (Bedrock, SageMaker, EKS/ECS, API Gateway, Lambda) with banking-grade network and security controls: VPC, PrivateLink, IAM, KMS, and logging and audit trails.
Architect hybrid deployments that connect cloud AI services with on-premise systems and self-hosted models (e.g., vLLM, Qwen) where data residency requires it.
Lead technology selection for vector databases, embedding and foundation models, orchestration frameworks and observability tools, with documented trade-offs on capability, cost, latency and compliance.
Engineering Excellence:
Build reusable foundations: retrieval templates, MCP server and tool-schema standards, guardrail middleware, evaluation harnesses and CI/CD for LLM applications.
Personally deliver the most complex components: orchestration core, complex retrieval, guardrail engine and legacy core-banking integrations.
Serve as the final escalation point for critical production issues, including hallucination, retrieval regression, latency spikes and tool-call failures. Lead post-mortems.
Governance and Stakeholder Leadership:
Define and own production KPIs: hallucination rate, retrieval recall@k, tool success rate, latency, availability and cost per request. Own the evaluation framework that gates deployment.
Represent AI engineering to Risk, Compliance, Information Security, IT Infrastructure and business units. Ensure auditability, explainability, traceability and alignment with the bank's AI governance and model risk policies.
Raise engineering standards through design reviews, code reviews and 1-on-1 mentoring.
REQUIREMENTS
Must have:
Minimum 7 years in software, ML or cloud engineering, including at least 3 years of hands-on production work on LLM, RAG or Agentic systems. At least 2 AI systems shipped to production with real users.
MANDATORY: proven AWS Solution Architect experience. You have designed and deployed AI services on AWS at production scale, including network and security architecture.
Expert-level Python (async, Pydantic, modular design) and deep hands-on experience with LangChain/LangGraph or an equivalent orchestration framework.
Production experience with vector databases (Qdrant, Milvus, OpenSearch or Pinecone), self-hosted LLM serving (vLLM, quantization, batching) and MCP or tool-calling design.
Experience designing LLM observability and evaluation (Langfuse, LangSmith or equivalent).
Strong architectural judgment on failure modes, blast radius, tail latency and cost. Excellent design documentation and the ability to defend decisions with data.
Bachelor's degree or higher in Computer Science, AI or a related field.
Preferred:
AWS Certified Solutions Architect - Professional, or AWS Machine Learning Specialty.
Experience in banking, financial services or other regulated industries.
Fine-tuning (LoRA/QLoRA) for Vietnamese or financial-domain language.
Production experience with OCR/VLM, voicebot or other multimodal AI.
Open-source contributions to AI tooling.
====================================================================
POSITION 2 - PRINCIPAL DATA SCIENTIST
====================================================================
KEY RESPONSIBILITIES
Analytics Strategy:
Define the bank-wide analytics strategy and use-case roadmap across business lines (retail, SME, cards, lending, deposits, bancassurance, wealth) and operational functions (collections, fraud, contact centre, branch network, cost-to-serve).
Identify, size and prioritise high-value opportunities (cross-sell and up-sell white space, retention, pricing, efficiency) with clear P&L impact and business cases.
Establish a value-measurement framework so that analytics impact is tracked in incremental revenue, cost and risk terms.
Advanced Modelling and Delivery:
Lead the design and delivery of complex customer and operational models: segmentation, propensity, next-best-offer and next-best-action, churn and attrition, CLV, uplift, forecasting, pricing elasticity and profitability.
Design rigorous experiments (A/B tests, control groups, causal inference) to measure incremental lift, and embed them in campaign and journey execution.
Own the full model lifecycle with data engineering and MLOps teams: framing, feature engineering, validation, deployment, monitoring and recalibration.
Ensure models comply with the bank's model risk management, data governance, privacy and documentation standards.
Executive Advisory and Capability Building:
Act as trusted advisor to business heads and senior management, delivering clear, quantified, decision-ready insight in Vietnamese and English.
Partner with Marketing, Product, Risk and Operations to move decisions from intuition to evidence.
Set analytical standards, build reusable data assets and mentor data scientists and analysts.
REQUIREMENTS
Must have:
Minimum 7 years of hands-on data science or advanced analytics, including at least 4 years in banking, fintech or financial services.
Proven track record of shaping analytics strategy for business or operational functions, with models in production that delivered measurable financial impact.
Expert SQL and Python (pandas, scikit-learn, gradient boosting); comfortable with large, complex production data.
Strong statistics and experiment design: hypothesis testing, sampling, control groups, causal inference and uplift modelling.
Deep understanding of banking products, customer lifecycle and P&L drivers. Starts from the business question, not the algorithm.
Excellent executive communication in Vietnamese and English.
Bachelor's or Master's degree in Data Science, Statistics, Mathematics, Computer Science, Economics or a related quantitative field.
Preferred:
Experience with recommendation systems, survival analysis or pricing optimisation.
Enterprise DWH (Oracle), Spark, Airflow, MLflow and MLOps practices.
CDP, marketing automation or campaign management platforms.
Application of LLM/GenAI in analytics.
Knowledge of core banking data structures and Vietnamese banking regulations.