About me

I am graduating in December 2026, seeking full time roles from January 2027, working on GenAI, LLMs, and agentic systems — preferably in GenAI safety, agentic security, red teaming for prompt injection and jailbreaks, or privacy-preserving machine learning. Research scientist, applied scientist, research engineer, software engineer and security engineer roles are all a good fit.

I am an MS CS student at UW-Madison. I am broadly interested in trustworthy machine learning, with a focus on the security and privacy of agentic systems. During my program I worked as a research assistant with Prof. Somesh Jha and published research on staged optimization for jailbreaking (ICLR’25), prompt injections that bypass KAD-based detectors (AISec@CCS’25) and prompt sanitization with formal privacy guarantees and high utility (NDSS’26). This summer, I am interning at Google, with the App & Ecosystem Trust team, where I red team production agentic crawlers.

I was previously a Research Fellow (pre-doctoral researcher) at Microsoft Research (2021-2023) with Dr. Nagarajan Natarajan and Dr. Jayashree Mohan, where I applied machine learning to problems in the area of systems, such as network path simulation (AAAI’23) and efficient inference in LLMs.

I graduated from IIIT Delhi in 2021, with a B.Tech. in computer science and applied mathematics, where I worked on adversarial robustness with Prof. Mayank Vatsa and Prof. Richa Singh.

Other (non-academic) things about me

I love reading and learning new things. This has led to many-a-fall down rabbit holes across the internet, and as a result, I have come to appreciate all sorts of niche topics. I have read enough sci-fi and fantasy that I feel compelled to craft my own setting and write short stories. I really like instrumental music, which mostly comes from having played and seen far too many games and movies with awesome scores. To that end, I learnt the tabla (indian percussion) a while back and picked up the violin more recently. I enjoy playing boardgames, especially partial information games like BoTC and Coup. I also occasionally bake breads from scratch 😄.

Recent News

• [May 2026]: I will be interning at Google this summer, in Mountain View, CA.
• [April 2026]: I passed my qualification exam! I am now a PhD Candidate.
• [August 2025]: Our paper titled "How Not to Detect Prompt Injections with an LLM" was accepted at AISec 2025! [Paper] [GitHub]
• [July 2025]: Our paper titled "Prϵϵmpt: Sanitizing Sensitive Prompts for LLMs" was accepted at NDSS 2026! [Paper] [GitHub]
• [January 2025]: Our paper titled "Functional Homotopy: Smoothing Discrete Optimization via Continuous Parameters for LLM Jailbreak Attacks" was accepted at ICLR 2025! [Paper] [OpenReview] [GitHub]
• [December 2023]: Our paper titled "Preempt: Sanitizing Sensitive Prompts for LLMs" was accepted at PPAI 2024! [Paper]
• [August 2023]: I was awarded the prestigious J.N. Tata Endowment scholarship for higher education.
• [July 2023]: I will be starting my PhD at UW-Madison in September! I plan to work on robustness and security in machine learning.
• [February 2023]: My co-author Sriram presented our paper “Simulating Network Paths with Recurrent Buffering Units,” during a poster session at AAAI 2023. [Poster]
• [November 2022]: Our paper titled “Simulating Network Paths with Recurrent Buffering Units,” was accepted at AAAI 2023! [Paper]
• [August 2021]: I started working as a Research Fellow (pre-doctoral researcher) at Microsoft Research India, with Dr. Nagarajan Natarajan.
• [July 2020]: Our paper titled “WaveTransform: Crafting Adversarial Examples via Input Decomposition,” was accepted at the AROW Workshop, ECCV 2020! [Paper]
• [May 2020]: Started my internship at TCS Research & Innovation on recommendation systems with Mr. Pankaj Malhotra.
• [February 2020]: I got accepted as a MITACS Globalink Research Intern at ETS Montreal, where I will work with Prof. Eric Granger on video surveillance.
• [August 2019]: Our paper titled “A Hybrid Approach for Tiger Re-Identification,” was accepted at the CVWC Workshop, ICCV 2019! [Paper]