Building intelligent systems and research-grade deep learning — from handwritten math recognizers to low-level systems programming.
I'm a Computer Engineering undergraduate at IOE Purwanchal Campus, Dharan — passionate about machine learning, deep learning systems, and applied AI research.
My work spans end-to-end ML pipelines: designing transformer architectures for handwritten math recognition, and building low-level systems in C and C++.
I'm driven by the intersection of rigorous engineering and real-world impact — building systems that don't just work in notebooks, but actually ship.
B.E. Computer Engineering · Dharan, Nepal
An end-to-end deep learning pipeline that converts images of handwritten math into LaTeX strings. Built a ResNet-34 backbone (truncated at layer3, stride 16) feeding 128 spatial tokens into a Transformer Encoder–Decoder. Trained from scratch on MathWriting (~230k samples). Live demo on HuggingFace Spaces.
Interactive drag-and-drop logic circuit simulator in C++ with SFML. Supports all fundamental gate ops, real-time signal toggling, and polished keyboard shortcuts.
GitHubReal-time multi-client TCP group chat in C using Winsock and POSIX threads. Server-client architecture with concurrent message handling and cross-device support.
GitHubResearch collaboration, project ideas, or just a hello — inbox is open.