AI Engineer / Design Engineer · LLM / RAG / multi-modal · Fullstack · Remote / hybrid / on-site
I ship AI systems end-to-end in production: corporate multi-LLM applications, RAG services, ASR post-processing, model evaluation frameworks, and Gen AI enablement programs at the department level. Worked across all generative-AI modalities — text / image / video / audio, node graphs, API orchestration.
In day-to-day work I use Claude Code, MCP, AI agents, and my own deterministic pipelines for automation. Fullstack: TypeScript / React + Python / FastAPI.
2.7 years as AI Engineer at Yandex, in AI since age 19. Also mentor at Yandex Practicum — Gen AI course for designers, 1y 6mo.
Currently building reframe — open-source AI-native engine for programmable design (see pinned ↓).
Yandex · AI Engineer · Sep 2023 — Apr 2026 · remote
- Multi-LLM orchestration app — TypeScript UI ↔ FastAPI ↔ provider aggregator. Daily use by content teams.
- RAG service for content automation — FAISS prototype → Qdrant production with web UI.
- AI Model Testing Framework — automated eval (chat / image / TTS / transcription / embeddings), VLM scoring, auto Excel reports.
- ASR pipeline + custom fact-loss benchmark (t1–t4) — solved LLM segmentation that broke quote boundaries via custom delimiter; validated across 3 models.
- Editorial content pipeline — source parsing → scaffold → LLM → humanizer → markdown / wiki autopublish.
- Gen AI enablement program (leading specialist) — motion / graphic design / illustration / presentations / web, including outsource teams.