David Gao// 高宇恒
whoami → LLM & AI Application Engineer
email david.gao.ai@example.comphone +86 136-8899-0011git github.com/davidgao-ailoc Beijing, China
summary
AI Application Engineer with 4+ years of experience in NLP and LLMs. Deep understanding of Transformer architectures, open-source model ecosystems (Qwen, Llama, DeepSeek), and RAG pipelines. Proven track record building enterprise AI agents and fine-tuned domain models.
stack
code
design
research
experience [2]
Senior LLM Research & Dev Engineer@ Cognitive Future AI Lab2023.03 - 至今
- Architected enterprise RAG system with Hybrid Search (BM25 + Dense) and BGE-Reranker, boosting QA precision from 78.5% to 94.8% and reducing hallucinations by 62%
- Developed multi-agent code generation workflow with LangGraph, automating task planning, coding, and unit test verification with 45% higher completion rate
- Deployed high-concurrency inference cluster via vLLM with PagedAttention & FP8 quantization, multiplying per-GPU throughput by 3.4x and cutting TTFT to 180ms
- Fine-tuned Qwen-72B for vertical domain using LoRA on 20k curated instruction pairs, outperforming base model by 28% on domain benchmarks
NLP Algorithm Engineer@ DeepData Intelligence2021.07 - 2023.02
- Trained entity extraction (NER) and sentiment classification models processing 10M+ daily documents for brand sentiment monitoring
- Built vector similarity retrieval system using BERT & Faiss, achieving sub-10ms nearest neighbor search over 10M vectors
projects [1]
OpenAgentic · Multi-Modal Agent Engine2023.09 - 2024.04
Python, LangChain, OpenAI API, Milvus, FastAPI
Open-source autonomous agent framework with browser automation, memory persistence, and tool invocation. 3.5k GitHub stars.
education
University of Chinese Academy of Sciences2018.09 - 2021.06
M.S. in Pattern Recognition & Intelligent Systems
Beihang University2014.09 - 2018.06
B.S. in Automation