Phát triển và duy trì data pipeline cho nguồn tri thức của ứng dụng LLM
Vận hành LLMOps: triển khai, giám sát, đánh giá và xử lý sự cố
Tóm tắt bằng AI từ JD gốc, đọc JD gốc trước khi ứng tuyển.
Join EPAM Vietnam as an AI Engineer to build advanced applications powered by Large Language Models. In this role, you will design, build and optimize applications powered by Large Language Models, with a strong focus on LLMOps, multi-agent systems and modern LLM integration patterns.
You will be part of an exciting initiative for a high-profile global client, collaborating with a diverse international engineering team to deliver next-generation AI solutions.
Responsibilities
Design and develop LLM-powered applications, including multi-agent systems (e.g., chatbots and agentic workflows) with prompt orchestration, tool/function calling and reliable multi-step coordination
Build retrieval-augmented generation (RAG) capabilities: data ingestion, chunking, embedding/indexing, vector search, reranking and citation/attribution to improve answer grounding
Develop and maintain data pipelines that curate and govern external knowledge sources used by LLM applications (ensuring freshness, quality and access control)
Improve model and system performance through prompt optimization, dataset curation, lightweight adaptation (e.g., LoRA/PEFT) and selective fine-tuning where appropriate
Implement and operate LLMOps practices: deployment, monitoring/observability, evaluation, incident response and iterative optimization for quality, latency and cost
Collaborate with ML engineers and data scientists to integrate LLM systems into existing products and services (APIs, platforms and downstream applications)
Track relevant advances in the LLM ecosystem and translate them into measurable improvements in our production applications
Requirements
Bachelor’s or Master’s degree in Computer Science, Mathematics, Data Science or a related field (or equivalent practical experience)
Strong proficiency in Python and experience building production-grade services (APIs, async jobs, testing and CI/CD)
Experience with LLM frameworks and tooling (e.g., LangChain/LangGraph, LlamaIndex) and modern LLM integration patterns
Proven experience in developing LLM applications and improving them via systematic experimentation (prompting, RAG tuning, dataset iteration and/or fine-tuning with PEFT)
Strong evaluation mindset with the ability to define success metrics and run offline/online evaluations (e.g., groundedness/faithfulness, relevance, retrieval metrics, win-rate, human review and A/B tests)
Experience in deploying models and LLM applications in production environments (containers, cloud services, scalable architectures and monitoring)
Familiarity with responsible AI and security practices (privacy, access control, prompt injection defenses and guardrails/policy-aligned behavior)
Clear communication and collaboration skills in English
