DmitryDmitriadi/ai-engineering-portfolio

AI Engineering Portfolio by Dmitry Dmitriadi: five production-oriented case studies solving “doesn’t stay alive in production” problems for LLM systems. Target users are engineers and teams building resilient, observable AI agents and automation (real-time conversational consultants, long-running multi-agent research, and browser automation). Key features: GraphRAG-based multi-agent workflows with reviewer-gated synthesis and stage graphs; a five-layer retrieval/memory stack (hybrid BM25+vector, source attribution, sub-100ms p95); a live AI consultant platform with sub-second turns and in-process cost guardians; resilient browser automation (self-healing locators, CDP shadow-DOM piercing, anti-detect); unified observability/evaluation (tracing across agents/tools, versioned error catalog, active cost enforcement). Tech highlights: LangGraph/OpenClaw, Qdrant, Playwright/CDP, OpenAI/Anthropic, Redis/Postgres, Docker. Contact: ddmitriadi@gmail.com.

AI-related keywords found in this repo's description, topics, or README summary — grouped by category. Each badge links to the corresponding ranking detail page.