I’m an undergraduate researcher at the PreCog Research Group, where I work on problems concerning AI safety and ML interpretability. I'm also passionate about competitive programming, 5-star rated on CodeChef, and expert rating on CodeForces. In my free time I love to read books on probability theory and statistics, research papers on problems I am working on or likely want to work in future.
I have worked on two research papers. One on enhancing AI safety using LoRA adapters and one on developing a data centric approach to AI safety, precisely generating better in distribution counterfactuals that help in improving model decision boundary. One interesting application of the latter is one can use it to augment their dataset with better counterfactuals thus use case can be extending to augment safety datapoints thus making AI more safe and robust. I can't link the papers, they are under review, considering ethical principles of conferences (COLING 2025), I am not allowed to do that right now.
One of the projects I am proud of are making a Network File Storage System in C, which supports concurrent interactions. I built this as course project of Operating system and networks taught by Dr. Karthik Vaidyanathan.
More projects that I made
Research Projects:
- Enchancing AI safety using LoRA adapters
- Work on mechanistic interpretability
- GEN-C (GEnerating Nearest Counterfactuals)
Internship:
- Full stack web app (Intelligent Wardrobe Management System.)
Course Projects:
- Drawing Editor HARPP(https://github.com/legend479/HARPP)
- Network File System in C(https://github.com/Harpreet287/Network-File-Sysyem)
Teaching: - Coding Tutorials on Lua game development at my university.
- Help in preparing lectures on Game Development with Unity for the course GDE at my university.