AI/ML engineer · b.tech cse (ai/ml) · lpu 2028
B.Tech CSE (AI/ML) at Lovely Professional University. I build and ship machine learning systems — NLP for low-resource Indian languages, vision models on edge hardware, and the plumbing that keeps them running.
fig. 01selected work
Each card carries the numbers it was judged on. The full set, with filters, is on the work page.
A fictional city modelled as a weighted graph, with a C++17 engine that routes emergency vehicles, spreads a disaster, computes evacuation capacity with max-flow and finds the roads whose loss cuts the network — every data structure written from scratch.
eventUniHack — product intelligence challenge
Turns a part number and a 35-character distributor description into a validated 252-column product record, with the evidence, method and confidence attached to every single value.
Binary particle swarm optimisation picks 7 features out of 30 on the Kaggle fraud dataset, and the honest result is a trade — better ranking metrics, slightly worse F1, at a quarter of the input.
fig. 03notes
Derivations rather than summaries. If I cannot explain where a formula comes from, I do not understand it yet.
currently
↑↓ move↵ open