Teams at Path

Develop and deploy production-grade computer-vision and vision-language pipelines to extract structured data from mobile-captured HMIS forms. Research and fine-tune VLMs and OCR/HTR models, optimize inference (batching, KV cache, quantization), design scalable cloud and on-prem GPU infrastructure, and enable multilingual, low-cost, edge-capable solutions for high-volume form processing.