Henil Panchal

Hi! I’m Henil Shalin Panchal, a final‑year Computer Science & Engineering student at Nirma University and a Research Intern at TCS Research, Physical Sciences Research Area, Mumbai.

In my role at TCS Research, I engineered advanced vision‑language models for ripeness and shelf‑life prediction in fruits and vegetables, built scalable inference pipelines with PyTorch and Flask, and delivered a cross‑platform Flutter app for end‑to‑end food tracking.

Previously, I served as an Undergraduate Researcher at the Daniel Kang Lab, UIUC, where I fine‑tuned phi‑series language models via LoRA, developed the LegalBench evaluation infrastructure, and adversarially tuned LLMs against prompt‑injection attacks using GCG and AutoDan.

Education & Skills

I’m completing my B.Tech CSE at Nirma University, Ahmedabad, with a CGPA of 8.3/10, excelling in Data Structures, Algorithms, Operating Systems, Big Data, and Software Engineering.

Projects & Open Source

On GitHub, I maintain 82 repositories with 198 stars, and my pinned projects include CodeExpert (AI assistant for coding), TinyStoriesGPT‑5M (5 M parameter Transformer), TeluguASR (Telugu speech recognition), and ChessSync (real‑time multiplayer chess).

My open‑source and academic contributions span deep learning, NLP, computer vision, and LLM security research.

Selected Publications

• Adaptive Attacks Break Defenses Against Indirect Prompt Injection Attacks on LLM Agents, NAACL 2025 Findings (arXiv:2503.00061)
• MMTEB: Massive Multilingual Text Embedding Benchmark, ICLR 2025 (arxiv:2502.13595)

Let’s Connect

Feel free to explore my work and get in touch—I’m always excited to collaborate on innovative AI and ML projects!