Abheesht Sharma

Bangalore, India

I’m Abheesht Sharma, an ML Engineer at Google working on the Keras ecosystem. I build and maintain KerasHub, KerasRS, and Tunix.

Previously, I was an Applied Scientist at Amazon, where I worked on ML techniques for advertisement fraud detection at scale. I hold a B.E. in Computer Science from BITS Pilani, Goa.

My research interests span NLP, recommender systems, and deep learning. I’m passionate about open-source ML tooling and making state-of-the-art models accessible to practitioners.

news

Feb 01, 2026	Blog post published: “Build and train a recommender system in 10 minutes using Keras and JAX” on the Google Developers Blog.
Sep 01, 2025	Released Tunix — a JAX-native library for LLM post-training. Blog post.

latest posts

Feb 01, 2026	Build a Recommender System in 10 Minutes with Keras and JAX

selected publications

KerasCV and KerasNLP: Vision and Language Power-Ups

Matthew Watson, Divyashree Shivakumar Sreepathihalli, Francois Chollet, and 14 more authors

Journal of Machine Learning Research, 2024

arXiv Bib

@article{watson2024kerascv,
  title = {KerasCV and KerasNLP: Vision and Language Power-Ups},
  author = {Watson, Matthew and Sreepathihalli, Divyashree Shivakumar and Chollet, Francois and Gorner, Martin and Sodhia, Kiranbir and Sampath, Ramesh and Patel, Tirth and Jin, Haifeng and Kovelamudi, Neel and Rasskin, Gabriel and Saadat, Samaneh and Wood, Luke and Qian, Chen and Bischof, Jonathan and Stenbit, Ian and Sharma, Abheesht and Mishra, Anshuman},
  journal = {Journal of Machine Learning Research},
  year = {2024},
}

AnEMIC: A Framework for Benchmarking ICD Coding Models

Juyong Kim, Abheesht Sharma, Suhas Shanbhogue, and 2 more authors

In EMNLP 2022 (System Demonstrations), 2022

Bib Code

@inproceedings{kim2022anemic,
  title = {AnEMIC: A Framework for Benchmarking ICD Coding Models},
  author = {Kim, Juyong and Sharma, Abheesht and Shanbhogue, Suhas and Weiss, Jeremy and Ravikumar, Pradeep},
  booktitle = {EMNLP 2022 (System Demonstrations)},
  year = {2022},
}

DRIFT: A Toolkit for Diachronic Analysis of Scientific Literature

Abheesht Sharma, Gunjan Chhablani, Harshit Pandey, and 1 more author

In EMNLP 2021 (System Demonstrations), 2021

Bib Code

@inproceedings{sharma2021drift,
  title = {DRIFT: A Toolkit for Diachronic Analysis of Scientific Literature},
  author = {Sharma, Abheesht and Chhablani, Gunjan and Pandey, Harshit and Patil, Rajaswa},
  booktitle = {EMNLP 2021 (System Demonstrations)},
  year = {2021},
}