.png){kind=logo}
Yupeng Lu
Mid-level Backend & Full-Stack Engineer specializing in distributed systems
Connect with Yupeng
Yupeng already has a relationship with Reval, so a warm intro from us gets a much better response than cold outreach.
Recommended
Already have an account?
About
Built a production internal RAG-based Q&A assistant at Huawei for ~4,000 engineers over a 12M-document Elasticsearch corpus, replacing link-only search with synthesized answers and achieving 87% user acceptance while keeping hallucinations under 0.4%. Pairs rigorous offline benchmarking (RAGAS, PR-gated F1 improvements) with human A/B testing and OpenTelemetry-based production monitoring, and also has strong Kubernetes/SRE experience orchestrating 50+ gRPC services with major MTTR and pager-fatigue reductions.
Experience
Education
Key Strengths
- Built and deployed internal LLM Q&A assistant for ~4,000 engineers on a 12M-document Elasticsearch index
- Maintained production hallucination/error rate under 0.4% using similarity gating and misuse prevention
- Established rigorous evaluation loop: 400-question benchmark + RAGAS metrics + human A/B tests + live feedback (87% acceptance)
- Enforced measurable quality improvements (≥1% F1 improvement required per PR)
- Designed hybrid retrieval at scale (BM25 + FAISS) with latency/cost benchmarking (P99, GPU cost)
- Improved reliability of 50+ gRPC services with circuit breakers and adaptive rate limiting; reduced MTTR 20 min → 3 min and pager fatigue by 70%
- Strong cross-functional delivery with legal/engineering leadership via weekly demos and dashboard-driven alignment
Discover more candidates like Yupeng
Search across thousands of pre-screened, high-quality, high-intent candidates on Reval.
Search Talent