Yanjie Zheng

LLM systems
that hold up in
production.

Retrieval, evals, and orchestration for enterprise workflows. Currently at Valdera.

Work

Valdera

Shipped Valdera’s first production AI system: an LLM supplier-discovery pipeline on Cloud Run replacing a manual workflow that took hours per request. Co-owned the V2 redesign on LangGraph.

Cut inference cost ~10x — $80 to $8 per request — by re-architecting on an agentic framework and trading retrieval volume for precision.

Owned the LLM platform layer. Migrated off Vertex AI to OpenAI behind a schema-validated wrapper: structured outputs, tool use, prompt caching, retry and backoff.

Built the eval harness and observability that made quality measurable, then cut model output reject rate from ~70% to ~30%.

Automated ~600 inbound emails a week of manual ops triage with an LLM classification and workflow system.

Replaced a third-party search vendor with a custom retrieval stack after coverage proved inadequate for Chinese chemical manufacturers.

Member of Technical Staff · 2025—26

Silicon Valley Commerce

Led AI from age 20. Hired and ran the team; two of them are now YC W26 founders.

Head of AI — acquired by FundPark · 2024—25

GenAI @ Berkeley

Grew to 60+ members, 23 industry projects, $135K revenue, and partnerships with Google, Netflix, and Rippling. Largest student-run GenAI organization in California.

Co-founder · 2023—25

Earlier: AI engineering intern at Google. ML assistant at Quizlet, where I built an internal LLM eval library adopted by 10+ engineers.

Contact

If something in your pipeline isn’t holding up, I’m happy to talk.