Palash Shah
Co-Founder at Bezel (YC W25)
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolPalash Shah is an engineer with eight years of hands-on experience building ML-driven products and developer tools, currently contributing at LangChain in New York. He co-founded a YC-backed startup focused on consumer marketing and previously led engineering work at Samsara, Zapier, and Roboflow, blending product-first thinking with production ML. His open-source work on "libra" highlights practical ML ergonomics—data preprocessing, CNN refinement, and TensorFlow/Keras model work—showing a strong bridge between research and applied engineering. Palash has repeatedly founded or helped scale early-stage ventures and infrastructure projects, bringing operational experience from shipping safety scoring, language modeling, and connected equipment systems. He pairs academic roots in computer science from UVA and bioinformatics research with a knack for making complex ML pipelines reliable in production. Not obviously visible from titles alone, he often sits at the intersection of founder-level product strategy and deep implementation detail.
8 years of coding experience
5 years of employment as a software developer
Westfield High SchoolY CombinatorBachelor's degree Computer Science, Bachelor's degree Computer Science at University of Virginia
Github Skills (11)
scikit-learn10
data-preprocessing10
neural-network10
mask-rcnn10
keras10
faster-rcnn10
machine-learning10
tensorflow10
python10
artificial-neural-networks10
scikit10
Programming languages (9)
Github contributions (5)
Palashio/libra
Apr 2020 - Jan 2022
Ergonomic machine learning for everyone.
Role in this project:
ML Engineer & Data Scientist
ML Engineer & Data ScientistContributions:11 releases, 629 commits, 190 PRs in 1 year 9 months
Contributions summary:Palash primarily contributed to the development and refinement of machine learning models within the `libra` repository. Their work included implementing and testing various machine learning queries, such as regression and classification models, using libraries like TensorFlow and Keras. They also focused on data preprocessing techniques, including one-hot encoding and standardization, to prepare the data for model training and testing. Additionally, the user made modifications to the CNN by adding multiple convolutional layers.
pythonmachine-learning-for-everyonedata-scienceneural-networksmachine-learning
Palashio/awesome-libra
Jul 2020 - Aug 2020
A curated list of all things related to libra.
Contributions:15 commits, 13 pushes, 1 branch in 22 days
libra
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.