I spend most of my time where abstractions end.
I'm a programmer learning about machine learning inference, CUDA, and deep learning. I'm currently building a lightweight machine learning library Nabla in C++ and a neural network inference engine using C++ and CUDA.
Previously a research intern at the IIT Bhilai Innovation and Technology Foundation, building deep learning models for a DST-funded project on healthcare diagnostics. Also active in competitive programming across AtCoder, Codeforces, and LeetCode.
A lightweight C++ ML library with modular components for linear algebra, dataset handling, preprocessing, evaluation metrics, and predictive models. Core matrix class, CSV dataset loaders, and linear regression, built with template metaprogramming and OOP.
A neural network inference engine in C++ and CUDA optimized by offloading GEMM and convolution to CUDA while keeping control flow on the CPU.
Contributed to the DST PRAYAS–funded project "Transforming Healthcare Access through AI-Driven Diagnostics Applications and Mobile Health Solutions," designing deep learning models and AI agents for medical data analysis and diagnostic decision support.
Out of 2,930 competing teams, 2025.
ICPC 2025
Open to ML, and software engineering roles.
lakshyaabaranwal@gmail.com