LabCAS - Laboratory Catalog and Archive Service
NASA JPL . NIH/NCI

This one is about over a decade long collaboration between JPL and NCI. JPL built the Data Archive for Cancer Biomarker Research under EDRN (Early Detection Research Network). The idea was to foster collaboration, research (and Machine Learning of course) on otherwise ill distributed and unverified data. The underlying technology was created on the footsteps of NASA Planetary Data System. The archive currently stores data (in multiple modalities) from over 40 Cancer research institutions across US.

​

I worked as the lead Data Engineer to design and develop their Data Access APIs, Data Processing Workflows, ML Service (Framework and APIs), LLM based Data Search (Question Answering system), and as a Researcher to understand how to make Biomarker related answers (given by the LLMs) more reliable.