I'm Devansh Bhardwaj, I recently graduated from IIT Roorkee, majoring in Electronics and Communication Engineering. My current work spans AI Security, Multi-Agent Systems, and Reinforcement Learning. I loved to do cool things with AI hence my interests are not limited to the aforementioned fields. I thrive on independently-driven projects and have led a lot of impactful works. With a long-term vision of leading AI research initiatives in India, I'm actively seeking full-time roles or research assistantships at top-tier labs where I can contribute meaningfully and grow toward a future PhD.
Currently, I am working along side Prof. Zhijing Jin on Causal AI and AI Safety. I am also working at Sarvam.ai as a Research Scientist in their Autonomous Agents team.
Previously, I worked on developing a generalized communication protocol for multi agent systems called, Agent Context Protocol, at Prof. Karthik Narsimhan's Lab. I worked as an ML Security Researcher at Repello AI, focusing on automated red-teaming of AI agents. I also worked on video prediction at Prof. Debasish Ghose's Lab at IISc Bangalore, under the supervision of Meenakshi Sarkar.
I served as the Secretary of the Data Science Group at IIT Roorkee, which has been the heart and soul of my ML journey.
Have a look at My CV to know more about me.