- Work model
- Hybrid
- Experience
- 2-5 years
- Employment
- Not specified
- Compensation
- Not disclosed
- Technology signal
- 10 tags
Technology context
10Parsed from the vacancy text; ordered by relevance to this role.
AILLMGenAIAzurePythonRAGBackendAI Solution EngineeringFastAPILangChain
Full listing
Role description
Join EPAM Vietnam as an AI Engineer (LLM) and help shape the future of AI-powered retail decision-making. You will design agent workflows, prompt and retrieval logic and production-ready services that turn data into grounded, traceable recommendations for business users. Working with cutting-edge AI technologies, structured business data and real-world retail challenges, you will play a key role in creating impactful, scalable and user-centric GenAI experiences that deliver measurable business value.
Apply today and be eligible for a one-month sign-on bonus.*
- Terms & Conditions apply .
Responsibilities
- Translate category management needs into LLM and GenAI use cases and workflows
- Design agent orchestration including prompt strategy, retrieval, tool calling and state handling
- Build back-end services and APIs for LLM applications using FastAPI and Python
- Integrate LLM providers such as Azure OpenAI and OpenAI with streaming, retries and cost controls
- Implement retrieval patterns with chunking, metadata, citations and relevance tuning
- Define evaluation approaches for relevance, grounding, safety, latency and cost
- Improve quality through testing, observability, feedback loops and iteration
- Partner with product, data, engineering and business stakeholders to ship usable features
Requirements
- Proven experience with Python back-end development, async patterns and FastAPI
- Hands-on work integrating Large Language Model (LLM) APIs such as Azure OpenAI or OpenAI
- Knowledge of LangChain or LangGraph for agent orchestration and tool workflows
- Expertise designing tool or function calling, from schemas to retries, confirmation flows and payload contracts
- Experience implementing Retrieval-Augmented Generation (RAG) with traceable sources and robust evaluation
- Ability to manage agent state through checkpointing, resumability and session handling
- Understanding of secure design for LLM applications, from secrets management to access control and prompt leakage prevention
- Clear communication with technical and business partners