Thierry Damiba
Developer Advocate & Data Scientist focused on vector search
San Francisco, California
Summary
Technical specialist in vector search and agentic retrieval: Thierry focuses on building retrieval and memory systems for AI agents, combining dense and sparse (hybrid) search, multimodal context, real‑time memory/upserts, and reranking to make agent workflows reliable and performant at scale. qdrant+1
Developer advocate and educator: He regularly presents workshops and talks (ODSC, Haystack, Neo4j), authors technical guides and Medium posts, and hosts/participates in podcasts and webinars to help practitioners adopt production‑grade vector search patterns. odsc+2
Experience building production and secure ML systems in government contexts: Prior work at SAIC included machine learning for the Department of Homeland Security with emphasis on classification and forecasting, informing a pragmatic approach to reliability, observability, and security for AI systems. thierrydamiba+1
Open source practitioner and tool builder: Maintains and publishes tooling and evaluation frameworks (Agentglass, MaxQ) and workshop code (qdrant_odsc25) to enable reproducible evaluation, agent observability, and practical demos for vector search and agentic systems. github+2
Work
Education
Projects
Writing
Building Performant, Scaled Agentic Vector Search with Qdrant
October 1, 2025A detailed guide on designing reliable agentic retrieval: real-time memory layers, multimodal and hybrid retrieval, advanced filtering, reranking, and production patterns to reduce latency and improve quality for agentic workflows.
Stand on the Shoulders of Giants
December 16, 2023A Medium essay about leveraging existing tools and platforms to streamline AI assistant and web application development, encouraging pragmatic reuse and learning in public.
Two Approaches to Helping AI Agents Use Your API (And Why You Need Both)
Explains complementary approaches to enabling AI coding agents to interact reliably with APIs, discussing tradeoffs and why combining techniques leads to more robust agent behavior.