Hi, I am Anubha

AI Safety | Post-Training | Enterprise AI | Alignment

I am a Senior Machine Learning Researcher at Bloomberg, working on post-training, alignment, and evaluation of large language models.

Building LLM systems for enterprise use cases, focusing on reliability, grounding, and safe deployment.

Interested in why models behave the way they do, especially around uncertainty, hallucinations, and privacy risks.

Graduated from Carnegie Mellon University’s Language Technologies Institute (LTI), advised by Dr. Graham Neubig and co-advised by Dr. Maarten Sap. Previously held a Research Assistantship with Dr. Sauvik Das.

I like building systems that connect research ideas to real-world deployment.

Organizations I have worked with :

SPUD Lab, Carnegie Mellon University

Media and Data Science Research Lab Adobe

MIDAS Lab

Conversational AI

Apple

R&D

NUS

Machine Learning Lab

Exciting News

[May 2024] My work done for protecting the user-side privacy associated with online self-disclosure through identification and abstraction using Language Models - Reducing Privacy Risks in Online Self-Disclosures with Language Models is accepted in ACL 2024, Thailand!
[March 2024] My work on exploring calibration in different prompting styles have been accepted in NAACL, Mexico City! - Program-Aided Reasoners (better) Know What They Know.
[Jan 2024] I have joined Bloomberg as a Senior ML Research Engineer!
[Nov 2023] My work on exploring calibration in different prompting styles is now on Arxiv! - Program-Aided Reasoners (better) Know What They Know. In this paper, we compare the calibration of Program Aided Language Models (PAL) and text-based Chain-of-thought (COT) prompting techniques over 5 datasets and 2 model types: LLaMA models and OpenAI models. Our results indicate that PAL leads to improved calibration in 75% of the instances.
[Nov 2023] My work at CMU in collaboration with Georgia Tech is now on Arxiv! - Reducing Privacy Risks in Online Self-Disclosures with Language Models. In this paper, we take the initiative to protect the user-side privacy associated with online self-disclosure through identification and abstraction using Language Models
[Oct 2023] My work done in Carnegie Mellon University : "Counting the Bugs in ChatGPT's Wugs: A Multilingual Investigation into the Morphological Capabilities of a Large Language Model" got accepted in EMNLP 2023, Singapore!
[May 2023] My work done in Carnegie Mellon University : "Multi-lingual and Multi-cultural Figurative Language Understanding" got accepted in ACL 2023, Toronto, CA!
[May 2023] I have started working in Apple Research as an AI - ML intern for this summer at New York City!
[Feb 2023] My research at Delhi Technological University was awarded "Premier" and "Commendable" Research Awards. (Click to see the paper!)
[Aug 2022] Joined School of Computer Science, Carnegie Mellon University for pursuing a Master's Degree!
[Feb 2022] Awarded the prestigious Vector AI scholarship, by Vector Institute of AI, Canada!