Summary
Praveen Venkatesh is a research scientist with 14 years of experience at the intersection of neuroscience, information theory, and machine learning, currently driving research at Upstart after a productive stint at the Allen Institute and Carnegie Mellon. He develops theoretically grounded metrics and fast estimators to reveal how information and noise shape computation in biological and artificial neural networks, with work featured as a NeurIPS Spotlight and in IEEE Transactions. Praveen has translated neuroscience insights into practical ML interventions—e.g., injecting structured noise to improve few-shot generalization in CNNs—and has collaborated closely with experimentalists to validate theories on high-dimensional spiking data. His background spans building novel EEG hardware and biomarkers for clinical outcomes to defining measures of fairness and unique information in AI, demonstrating a rare ability to move ideas from theory through implementation to real-world validation. Notably, he achieves large computational speed-ups (1000×) on high-dimensional estimators, enabling analyses at scales previously infeasible.
14 years of coding experience
11 years of employment as a software developer
Indian Institute of Technology Madras
Doctor of Philosophy - PhD, Electrical and Computer Engineering, Doctor of Philosophy - PhD, Electrical and Computer Engineering at Carnegie Mellon University