Описание

Architecture of the agent: routing questions between sources, parsing composite requests, tool calling, context and session management.
RAG on regulatory documents and documentation: loading and chunking large documents, hybrid search, reranking and updating data when legislation changes.
Quality assessment (evals): forming reference question sets, automated testing before releases, and monitoring quality metrics in production.
Security and data: access control, protection against prompt injection, confirmation of write operations, audit of actions, and compliance with RK legislation on personal data.
Model and infrastructure selection: working with cloud APIs and on-premise solutions, optimizing cost and speed, monitoring and tracing requests.
AI functionality development: creating reports, filling out forms, and automating operation sequences.
Mentoring and code review: code and prompt review, developing technical competencies of the second developer.
Product development: assessing technical feasibility of ideas with the Product Owner, conducting technical demos, and interacting with customers.

From 3 years of experience in backend development, of which at least 1.5 years is commercial experience developing LLM products in production.
Practical experience implementing RAG: chunking, embeddings, vector databases, hybrid search, and reranking.
Experience developing AI agents, tool calling, API integration, and structured output.
Experience building evals, LLM-as-judge, and regression testing.
Proficient in Python, experience with APIs, queues, Docker, CI/CD, logging, and testing.
Knowledge of SQL and experience with PostgreSQL.
Understanding of LLM risks: prompt injection, data leakage, hallucinations, and methods to prevent them.
Experience in mentoring or conducting code reviews.

Senior AI Engineer