Soorya Narayanan

Machine Learning | Research | Production

I'm a Data Engineer and ML practitioner pursuing my MS in Applied Data Science at USC, building custom Autoencoder for Verizon funded research. I ship production ETL pipelines and ML systems that put models in front of real users.

Before USC, I owned wild face-recognition system boosting accuarcy from 42% to 93%, and cut lead-to-conversion time by 71%. I like problems with a number attached, if I can't measure whether it got better, I get a little suspicious of the win.

Poke around my projects, or grab the full CV to know more about me.

CURRENTLY BUILDING
  • PMNet (Verizon) — 98% fidelity, 19K+ samples
  • Immuny — clinical AI allergy tracker, in production
  • Intelligent Bistro — 236 tests, 94% coverage