Kalpan Mukherjee

In the final year of my master's in Computer Science at the Courant Institute of Mathematical Sciences, New York University, I am pursuing research primarily revolving around the recent growth in Large Language Models.

1. With the Emerge lab - I am working on exploring code evolution through LMs for building rule-based agents that perform on par with neural network-based RL agents in environments. A breakthrough in this field will allow agents to be more interpretable and amenable.
2. With the SAI lab - I am working on speeding up LLM inference through cached self-speculative decoding while using parameter-efficient fine-tuning techniques.
3. With my AWS industry collaboration - I am building a training-cum-inference recipe that allows one to interleavingly use Language Models and Neural Networks in tandem to achieve the highest possible inference accuracy for a task.

Previously, I completed my Bachelor's Degree in Information Sciences at Nitte Meenakshi Institute of Technology (NMIT), Visvesvaraya Technological University. My previous experience includes building an LM based agentic framework for database query regression analysis at AWS Redshift, buidling natural language and computer vision enabled backend systems as a Software Developer at Hashedin by Deloitte, developing a virtual classroom experience as a SWE Intern at Creatist and conducting research on Video Shot Detection in low illumination conditions at the Center for Robotics Research, NMIT.

Email  /  CV  /  LinkedIn  /  GitHub  /  Medium  /  YouTube