CS Honours Student | ML Researcher at ISI Kolkata | AI & Data Science





Bachelor of Science - BS, Computer Science and Statistics
Christ University, Bangalore

ISC, Commerce
Calcutta Boys' School
Join Md and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal BossCS Honours Student | ML Researcher at ISI Kolkata | AI & Data Science
I completed my 12th grade at Calcutta Boy's School with a score of 93.75% in ISC, achieving a perfect score in Computer Science. Currently, I’m pursuing a dual major in BSc Computer Science and Statistics at Christ University, Bangalore, where my passion for data science started taking shape. As I explored data analysis through tools like MS Excel and R, I realized the transformative potential of data in decision-making and innovation. I’ve since gained experience in data visualization using Tableau and am currently expanding my skills in Python, diving into its applications in machine learning and beyond. Driven by a commitment to harness the power of data, I’m eager to build intelligent solutions and make an impact in the field of data science. Let’s connect and share insights!
Greater Bengaluru Area
claude code
claude
notebooklm
openai
Research Intern at Indian Statistical Institute, Kolkata, Feb 2026 - Present · 8 mos, Designed and built TPMC (Temporal Prompt-Aware Memory Consolidation), an online machine-unlearning framework that clusters streaming data into temporally-cohesive prompt-keyed partitions and proactively deletes stale clusters via a novel Three-Gate criterion (relevance, population, uncertainty), achieving verified exact deletion (L1 = 0) at O(1) cost versus O(N) for full retraining. Redesigned the influence-scoring pipeline by replacing a degenerate Sinkhorn-OT approach with Sliced Wasserstein Distance, correctly capturing distributional drift in high-dimensional (512–768-dim) backbone feature spaces and delivering a 33.56% accuracy gain over EWC and 28.89% over multi-head baselines on the HuffPost benchmark. Diagnosed and documented the method’s boundary conditions through rigorous ablations across four datasets (HuffPost, Yearbook, MNIST-Rotation, Synthetic Moons), producing rigorous negative findings that sharpened the paper’s scope for submission to ECAI 2026 / AAAI 2027.
Bachelor of Science - BS, Computer Science and Statistics, Christ University, Bangalore
ISC, Commerce, Calcutta Boys' School
github: https://github.com/kamran-rashid