Hua Li

Python, Java/Spring Boot, SQL · TensorFlow/Keras, scikit-learn · Elasticsearch/OpenSearch, Lucene · ChromaDB, sentence-transformers · Kafka, Postgres, AWS, Docker

Hybrid retrieval and cross-encoder reranking. Also reading up on how people are evaluating LLM-as-judge pipelines, since my own results depend on it more than I expected.

Two side projects: a deployed RAG system with its own evaluation harness, and a neural rebuild of a TREC-ranked recommender.

Conversations about retrieval evaluation, RAG that works beyond the demo, and recommender systems. Happy to talk through the experiments in my posts.