ai engineer

EPAM is a provider of digital transformation services, product engineering, and digital solutions, helping clients build and scale intelligent applications powered by modern technologies.

• Design and build async endpoints, background tasks, and streaming responses using Python and FastAPI;
• Develop and maintain agent-based systems, including prompt design, tool-calling, and RAG pipelines;
• Diagnose and resolve LLM failure modes such as hallucination, context contamination, and non-deterministic output;
• Tune retrieval quality through chunking strategies, embeddings, and vector search optimization;
• Build and fix production-grade React UIs, including layout, responsive behavior, and streaming state management;
• Identify and resolve concurrency issues, including shared mutable state and callback bleed under concurrent requests;
• Operate within Docker and Kubernetes environments, including deployment and pod/log debugging;
• Write and maintain test coverage using pytest and vitest, including mocks and fixtures;
• Trace symptoms to root causes across frontend, backend, and agent layers to ensure system reliability.

• 5+ Years of experience with Python, FastAPI, and asynchronous programming patterns;
• Proficiency in agent frameworks such as pydantic-ai, LangGraph, or LangChain, with hands-on experience in prompt engineering and tool-calling;
• Expertise in RAG systems, including chunking, embeddings, and vector search tools such as Qdrant or pgvector;
• Skills in React (hooks) and CSS, with the ability to build and troubleshoot complex, responsive UIs;
• Understanding of concurrency and state management, with experience implementing request-scoped state, locks, and thread-safety;
• Familiarity with Docker and Kubernetes, including container management and deployment debugging;
• Background in testing practices using pytest and vitest, with consistent use of mocks and fixtures;
• Strong debugging skills with the ability to trace issues across frontend, backend, and agent layers;
• English proficiency at B2 level or higher;
• Nice to have: Experience with AWS Bedrock or other model gateway platforms beyond OpenAI, Familiarity with Langfuse or other LLM observability tooling, Knowledge of python-pptx or document generation tools for PowerPoint and PDF outputs, Background in regulated industries such as pharma or life sciences, including familiarity with FDA PMA/510(k) and IVDR terminology, Skills in Helm chart authoring and CI/CD pipelines using GitHub Actions, Design sensibility for building premium, non-generic user interfaces, Experience integrating SSO/OAuth solutions such as Azure AD or MSAL, Familiarity with vector DB operations, including Qdrant collection management and embedding pipeline maintenance, Understanding of prompt-eval or LLM-testing frameworks for measuring answer quality over time.

• No conditions specified