Avanti Shrikumar
Postdoctoral Researcher at Stanford University
Palo Alto, California, United States
Summary
👤
Senior🎓
Top SchoolAvanti Shrikumar is a computational biologist-turned-earth scientist who blends 11 years of software and research experience to apply machine learning to biological and oceanographic problems. After a CS PhD at Stanford working on regulatory genomics and model interpretability in the Kundaje lab, she pivoted to study oceanic nutrient cycling as a postdoctoral researcher with the Casciotti lab and Stanford Data Science Institute. She is a hands-on ML engineer and back-end developer, contributing to influential open-source tools such as the deeplift library to improve model explanation and multimodal input support. Her background ranges from high-throughput bioinformatics at MIT to pragmatic systems engineering as a Forward Deployed Engineer at Palantir, which informs both her reproducible research practices and production-minded code. Based in Palo Alto, she maintains active codebases under kundajelab and nitrogenlab on GitHub and combines deep algorithmic insight with domain-driven science. Colleagues value her ability to move between rigorous methodological development and real-world environmental questions.
11 years of coding experience
4 years of employment as a software developer
Bachelor of Science (BS), Computational Biology, 5.0/5.0, Bachelor of Science (BS), Computational Biology, 5.0/5.0 at Massachusetts Institute of Technology
Spanish, Hindi, English
Github Skills (7)
algorithm10
algorithms10
machine-learning10
deep-learning10
python10
theano10
tensorflow9
Programming languages (5)
Github contributions (5)
kundajelab/deeplift
Apr 2016 - Nov 2021
Public facing deeplift repo
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:21 releases, 916 commits, 67 PRs in 5 years 7 months
Contributions summary:Avanti contributed to the deeplift repository, focusing on machine learning model interpretability. Their work involved the implementation of the "revealcancel-redist" as a new mode, indicating involvement in developing and refining the algorithmic approach for DeepLIFT. The user also contributed to supporting multimodal inputs and various fixes related to pre-existing implementations, demonstrating an active role in both the core architecture and practical functionality of the library. The user's code changes were geared towards enhancing the library to support a broader range of machine-learning scenarios.
pythoninterpretable-deep-learninginterpretabilityintegrated-gradientssaliency-map
kundajelab/simdna
Jul 2016 - Sep 2021
A python library for creating simulated regulatory DNA sequences
Contributions:4 releases, 129 commits, 12 PRs in 5 years 2 months
python-librarypythonregulatorydnadna-sequences
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.