About
What I do
I work on production ML on noisy real-world sensor data. Co-built RoboSat at Mapbox. Led data science at Rapido on ride-hailing marketplace pricing and forecasting. Earlier work at Yulu, NextDrop, and Mindshift on sensor and fleet platforms.
The shape of the work has been consistent: extract structured signal from messy physical-world data, make operational decisions from it, close the loop with feedback.
Based in Bengaluru.
Where I’m going
Moving toward neurotech and biosignal applications. The same engineering pattern, with biology on the input side: extract structured signal from noisy neural and physiological data, make decisions or generate outputs from it.
Areas I’m exploring:
- Mental-health wearables and digital therapeutics — physiological signal → state inference → intervention
- Clinical EEG and biosignal platforms — signal quality, artifact rejection, closed-loop neuromodulation
- Brain-computer interfaces — neural decoding, motor imagery, non-invasive and invasive
- Neuro-AI and computational neuroscience — population activity modeling, decoding latent state
Completed g.tec Spring School of NeuroTech 2026. Currently building public neurotech work — entries land on the Projects page as they ship.
Contact
- Email — bhargav.kowshik@gmail.com
- LinkedIn — linkedin.com/in/bkowshik
- GitHub — github.com/bkowshik
- X — @bkowshik
Always happy to talk about biosignal, neurotech, or ML on noisy real-world data.