Keshav Santhanam - Applied Deep Learning Research Scientist at NVIDIA

Keshav Santhanam

Applied Deep Learning Research Scientist at NVIDIA

Cupertino, California, United States

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Keshav Santhanam is an applied deep learning research scientist with a decade of experience building and optimizing ML systems, currently at NVIDIA after a long research tenure at Stanford. His work spans systems and models—from enabling CPU execution and C++ extensions for the state-of-the-art ColBERT neural search to improving transparency in language model evaluation frameworks like HELM. He combines systems-level optimizations for distributed and heterogeneous training with practical engineering (IRs for auto-distribution, schedulers for GPU clusters) informed by multiple internships at Microsoft and Google. Based in Cupertino, he brings a PhD-level research background and a track record of shipping reproducible, production-oriented contributions that bridge research code and real-world performance. An under-the-hood detail: he has repeatedly focused on making GPU-first ML research usable on CPU and mixed environments, improving accessibility and deployment fidelity.

10 years of coding experience

9 years of employment as a software developer

Bachelor's degree, Computer Science, 3.86, Bachelor's degree, Computer Science, 3.86 at University of Illinois Urbana-Champaign

Monta Vista High School

Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Stanford University

English, Spanish

Github Skills (15)

pytorch10

machine-learning10

javascript10

openai-api10

api10

optimisation10

python10

apidoc10

ant10

optimization10

front-end-development9

cprogramming-language9

nlp9

c-language9

cuda5

Programming languages (3)

C++Jupyter NotebookPython

Github contributions (5)

stanford-futuredata/ColBERT

Jul 2021 - Jan 2023

ColBERT: state-of-the-art neural search (SIGIR'20, TACL'21, NeurIPS'21, NAACL'22, CIKM'22, ACL'23, EMNLP'23)

Role in this project:

ML Engineer

Contributions:18 reviews, 77 commits, 24 PRs in 1 year 5 months

Contributions summary:Keshav primarily focused on enhancing the ColBERT neural search model by adding CPU execution support. They modified core components, including candidate generation and residual codec implementations, to enable CPU-based processing. These changes involved conditional logic to handle GPU/CPU execution paths, optimizing the algorithms to work efficiently without a GPU. Further contributions included integration of C++ extensions to improve performance.

pytorchnaaclartneuripsstate-of-the-art

stanford-crfm/helm

Feb 2022 - Sep 2022

Holistic Evaluation of Language Models (HELM), a framework to increase the transparency of language models (https://arxiv.org/abs/2211.09110). This framework is also used to evaluate text-to-image models in HEIM (https://arxiv.org/abs/2311.04287) and vision-language models in VHELM (https://arxiv.org/abs/2410.07112).

Role in this project:

Back-end Developer

Contributions:35 commits in 7 months

Contributions summary:Keshav primarily focused on improving the language model evaluation framework. They implemented the parsing and display of finish reasons in the UI, enhancing the user's ability to understand model behavior. Key contributions include modifications to the OpenAI, AI21, and Anthropic client code to incorporate finish reason information. They also added metrics related to finish reasons and corrected various formatting issues within the codebase.

nlparxivabsberthelm

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.

Request Free Trial