~/yanfei $ San Jose, CA · --:--:-- · 37.34° N, 121.89° W
status
$ ▍
> now : agentic AI & ranking @ PayPal
> prev : LLM serving & post-training @ NetMind.AI
> else : papers · mountaineering · astrophotography
$ ▍
// fig. 02 — selected work
Things I've built
// fig. 03 — toolkit
The craft
## Recommendation / Retrieval
- Two-Tower / Dual-Encoder Retrieval
- Ranking · CTR / CVR Prediction
- Learning-to-Rank
- DCN · DLRM · Embeddings
- ANN · FAISS · HNSW
- A/B Testing · recall@K · NDCG
## LLM / Agentic AI
- Multi-Agent Orchestration
- Tool Use · MCP
- RAG
- Fine-tuning · PEFT / LoRA
- SFT · RLHF · Post-training
- Prompt Engineering · Evals
## ML Systems / Serving
- LLM Serving · Inference Optimization
- Quantization · TensorRT
- DeepSpeed · SGLang · vLLM
- Distributed / Multi-GPU Training
- Latency & Throughput Optimization
## Frameworks & Tools
- Python · SQL
- PyTorch · TensorFlow
- LightGBM · XGBoost
- Spark · Feature Engineering
- Git · Docker · CI/CD
// fig. 04 — field notes
Through my lens
drag to explore
// coda
Let's make something worth shipping.
Open to conversations about recommendation systems, agentic AI, or where to catch the best alpenglow.
open to opportunities