Fail-closed observability
If your sanitizer fails, the write should fail with it. Healthcare AI traces are only useful when the safety path is as reliable as the happy path.
Ruchir
Kalokhe
Backend engineer leaning into AI research.
Love AI — building with it, and building it.
Projects
AuscultPrivacy-preserving observability for healthcare AI agents
TraceLinkManufacturing supply-chain traceability
Notes
Ranking without keyword stuffing
Keyword stuffing still wins naive retrieval. Structured fit scoring and evidence-guided expansion change what “top candidate” actually means.
Research models as services
The interesting work is rarely the model alone. Health checks, packaging, and cutting redundant detection often matter more than another parameter tweak.