This is Ashima, a PhD student at UCLA CS advised by Prof. Saadia Gabriel.

I actively collaborate with Prof. Nanyun Peng, Prof. Kai-Wei Chang and Prof. Tom Hartvigsen. Additionally, I work with datacomp for data curation with Prof. Ludwig Schmidt.

My research focuses on post-training data recipes and training paradigms for improving the reasoning capabilities, alignment and safety of LLMs. I am currently exploring recipes to develop general-purpose reasoning models and software engineering agents.

Awards
✨ UCLA Computer Science Fellowship 2024
✨ Google Deepmind Scholarship 2021-2023 (UCLA MSCS)
✨ Mitacs Globalink Research 2019

News


🍄 July 2025 The preprint for my first paper as a PhD student is out! ModelCitizens: Representing Community Voices in Online Safety.

🍄 June 2025 OpenThoughts preprint is out! We release the OpenThinker-7B, the new SOTA open-data 7B reasoning model!

🍄 June 2025 AI Debate Aids Assessment of Controversial Claims preprint is out!

🍄 May 2025 Joint Preference Optimization got accepted to ACL Findings 2025!

🌻 Checkout SheFrames, a crowdsourced art project celebrating womanhood. Submit your stories here.


Service


  • Reviewer: ICLR Tiny Papers 2024, SoCal NLP Symposium 2023, Neurips D&B 2024, COLM 2024, COLM 2025, ACL May ARR 2025
  • Organizer: UCLA NLP Seminar Series