Akshay Balsubramani
Lead at On Deck
San Francisco Bay Area United States
Summary
👤
Senior🎓
Top SchoolAkshay Balsubramani is a lead AI scientist and entrepreneur in the San Francisco Bay Area with 11 years of experience applying machine learning to drug discovery, genomics, and chemistry. He builds bespoke AI agents and tooling that let interdisciplinary teams—chemists, biologists, and executives—interact with data in real time to accelerate decisions across discovery-to-clinic pipelines. His work spans fundamental ML research (PhD-level theory and single-cell genomics at Stanford) to productized impact, including saving 40% of reported leads in a late-stage protein campaign and launching multiple monoclonal antibody programs. At VAC and as a consultant to venture and pharma, he’s combined generative models, ADMET/synthesis prediction tools, and visualization systems to scale team workflows. He also contributes code to influential open benchmarks like WILDS, adding models and data loaders for genomics tasks, reflecting a commitment to robust, reproducible ML for real-world distribution shifts. Akshay’s background blends rigorous theory, hands-on engineering, and domain fluency in biology—enabling unusual velocity from prototype models to experimental proof-of-concept.
12 years of coding experience
11 years of employment as a software developer
EECS, EECS at University of California, BerkeleyUniversity of California, San Diego
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (11)
data-preprocessing10
data-loading10
pytorch10
machine-learning10
deep-learning10
dataprep10
preprocessing10
python10
load-data10
preprocess10
tensorflow4
Programming languages (4)
Github contributions (5)
p-lambda/wilds
Feb 2021 - Jul 2021
A machine learning benchmark of in-the-wild distribution shifts, with data loaders, evaluators, and default models.
Role in this project:
ML Engineer
ML EngineerContributions:112 commits, 76 pushes, 3 branches in 5 months
Contributions summary:Akshay implemented and added new models and data loaders to the `wilds` repository, a machine learning benchmark for in-the-wild distribution shifts. They added rough cut models and data, including files for specific datasets like `EncodeTFBSDataset` and model files such as `CNN_genome.py`. Their work involved modifying existing code files to integrate new models and improve the dataset fetching and preprocessing functionalities of the project.
dataloadermachine-learning
b-akshay/pc-autoencoder
Nov 2016 - Feb 2022
Contributions:12 commits, 9 pushes, 2 branches in 5 years 4 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.