artem = {
    "role": "AI Architect · Founding AI Engineer",
    "built": "the AI core of an enterprise GenAI platform",
    "focus": ["RAG", "AI agents", "LLM evaluation"],
    "base": "Python backend: FastAPI, microservices, queues, cloud",
    "domain": "regulated industries",
    "range": "architecture → code → production",
    "coding_since": 2019,
}

• Architected and built the AI core of an enterprise GenAI platform: multi-pass multimodal retrieval, LLM orchestration, and per-stage model tiering — co-designing the architecture at CTO level.
• As the product's first engineer, took it from an empty repo to live enterprise demos. Those demos brought the product its first enterprise contracts.
• Built a pipeline that compresses a week of manual financial analysis into about a day — every number recalculated from the source documents and cited.
• Built the evaluation layer a regulated product needs before it can go live: F1 scoring across text, tables, and images, semantic image similarity, and hallucination detection wired into an automatic improvement loop.
• Run the engineering around it: the Python backend team, the release scopes, and the standards every PR is held to.
• Earlier: drove GenAI adoption across a global consultancy — agent design patterns at an internal AI education day, backend response times cut by up to 50% — and gave a university guest lecture on AI agents.

• Typed, structured model output everywhere. No hand-parsing, no silent fallbacks.
• Metrics are computed in code, never by the model.
• Model tier and reasoning effort are chosen per pipeline stage, on cost and accuracy trade-offs.
• A missing value is reported as "not found", never guessed.
• The dev loop is part of the system: role-scoped subagents, multi-angle code review with an auditor pass, deterministic hooks that keep secrets out of reach, and tracing across every LLM call.

AI / LLM

Languages & frameworks

Data & messaging

Cloud & DevOps

• Generative and Agentic AI Programme — Saïd Business School, University of Oxford (2026)
• AWS Certified AI Practitioner — in progress