SUMMARY
AI Engineer & Solutions Architect who turns fuzzy ideas into reliable, production-grade systems. My focus is end-to-end: data ingestion and ETL, retrieval (vector + traditional search), LLM/RAG and agent workflows, APIs, and the cloud/DevOps glue that keeps it all running. I build modular services and pragmatic architectures, lead and mentor small teams, and keep reliability, observability, security, and cost control at the core. Experience spans knowledge assistants, internal search, analytics platforms, and complex integrations across multiple industries. Especially interested in building and leading AI platform work, modernizing legacy backends and data flows, and partnering on thoughtful, long-term products.
EXPERTISE
EXPERIENCE
LEAD_AI_ENGINEER
[AGENTIC_LAB]
- → Built AI assistants for large enterprises on private infrastructure — vLLM inference, LangGraph stack
- → Intent detection, parallel tool fan-out, planning mode, subagents, streaming, context-engineering middleware
- → Tuned hybrid vector + lexical retrieval, lifting precision@k and recall@k
- → Designed a DWH for a ~500k-document legal corpus with LLM-assisted ETL pipelines
- → Led an AI-lab team (5+); pioneered observability- and evaluation-first delivery with LLM-as-judge
LEAD_AI_ENGINEER / SA
[SOLO GPT]
- → Designed microservices-based ETL architecture, DWH orchestration, knowledge base for sales assistant
- → Served top-5 Russian edtech, peak traffic 1000+ RPS
- → Standardized CI/CD; prepared Kubernetes migration
- → Architected integration hub using Temporal and MCP
FREELANCE / INDEPENDENT
[SELF_EMPLOYED]
- → Technical advisor for web3 startup
- → Built corporate chat assistant using Next.js, Anthropic SDK
- → Combined vector search with SQL queries
TECH_LEAD / SA
[RED_MAD_ROBOT]
- → Designed SaaS and on-prem LLM solutions
- → RAG platform, agent workflows, hybrid vector search
- → Supported hiring and mentoring
BACKEND DEV / TECH_LEAD
[INSOMNIAC_DESIGN]
- → Carbon footprint analytics for Adidas, Apple, Mercedes-Benz
- → Managed international teams 5-10 engineers
- → Achieved 100× search speed-up
FULL_STACK_DEVELOPER
[FUTURE_FERTILITY]
- → AI platform for embryologists
- → Optimized ML container resources
- → Built analytics dashboards in Apache Superset
QUANTITATIVE_DEVELOPER
[SBERBANK_CIB]
- → High-performance APIs for trading panel
- → Designed ETL for critical market decisions
EDUCATION
MOSCOW_STATE_UNIVERSITY
Bachelor's degree, Journalism
2013 → 2018