Software engineer focused on search systems, algorithms, performance optimization, and machine learning infrastructure.
Founder of Software.Land, building systems for semantic knowledge representation, search, and learning progression.
Source-available search engine built with TypeScript, featuring positional inverted indexes, compact typed-array representations, block-level pruning, and ranking optimizations.
Contributions to the C/C++ LLM inference ecosystem:
- #29060 — HTTP 400 handling for invalid embedding requests
- #27920 — Qwen3.5 embedding model conversion support
- #16541 — Raw embedding output support
Search and information retrieval · Algorithms and data structures · Distributed systems · LLM inference · Knowledge graphs