Trinav Chaudhuri
Applied AI / ML Engineer
I bulid AI that works when there is a consequence to being wrong.
I care about the little things. Pipelines, retrieval, evals, the research that decides whether a model is useful or just fluent. Unglamorous, but where most of my time is spent.
I started uni as a pre-med, but couldn't stay away from tech. Now, I'm working towards combining my experience from both fields to build health AI..
Work
All roles- AI/ML EngineerOmi Health
Clinical AI infrastructure, Retrieval grounded in real interventions, note generation, evaluation harnesses, and the agentic workflows that link them.
Feb 2026 – now - Data Science InternMagna Education
Cheating detection over quiz answers: embeddings and similarity clustering against a labeled benchmark of model-written and human-written answers, compared against two commercial detectors.
Jun 2025 – Aug 2025
Where I studied
My Path- M.S. Information Systems, Concentration in Health TechCornell UniversityAug 2026 – May 2028
- M.S. Applied Information ScienceTechnion – Israel Institute of TechnologyAug 2026 – May 2028
Taken concurrently as dual-degree program at the Jacobs Technion–Cornell Institute at Cornell Tech.
- B.A. Data ScienceUniversity of California, Berkeley2022 – 2026
- B.A. Molecular and Cell BiologyUniversity of California, Berkeley2022 – 2026
Double major.
Selected projects
All projectsAreWeThroughSports
A World Cup qualification tracker that refuses to publish when its two sources disagree. Built test-first, then checked against an independent simulation to see whether I had actually understood the maths.
RegWatchCompliance
Four agents that read a codebase for HIPAA violations. Static analysis handles what is provable, a language model handles what is only arguable, and the split between those two is the whole design.
EazyHealthHealth AI
Weekly health briefings and on-demand plain-language explainers at adjustable reading levels, sourced from CDC, NIH and Mayo, for people who were never going to read the original.
Writing
ArchiveCurrently
Reading
A mix of genres, currently exploring Japanese authors
Listening
Eighteen years of tabla has led me to appreciate rhythms over melodies
Everywhere I have been so far, on one slowly turning ball. Open the globe