I build software that is reliable, maintainable, and designed to evolve.
My work centers on backend engineering, developer tooling, and full‑stack systems — with a growing focus on distributed systems, systems programming, and applied machine learning. I care more about thoughtful architecture than clever complexity. Software should be easy to reason about, simple to maintain, and ready to scale when it needs to.
The workspace, tools, and technologies behind my daily engineering workflow.
Building scalable backend systems, developer tooling, and modern full-stack applications.
Explore My Featured Projects
Confidence-Aware Cheque Validation
An uncertainty-aware deep learning system for automated cheque verification, designed to improve prediction confidence and reduce false decisions through reliable model inference.
Next.js · FastAPI · PyTorch · MongoDB
Data-Driven Campus Operations
A full-stack canteen management platform that streamlines daily operations while integrating predictive analytics through scalable backend services.
Next.js · FastAPI · Python · MongoDB
Developer Tooling
A lightweight command-line utility for executing Python scripts and Jupyter notebooks through a Go-based runtime.
Go · Python
High-Performance Text Comparison
A plagiarism detection engine built in Rust using the Myers Difference Algorithm for efficient and accurate textual comparison.
Rust · Myers Diff
Real-Time Auction Platform
An asynchronous auction platform supporting concurrent bidding through persistent WebSocket connections and event-driven communication.
Python · AsyncIO · WebSockets
Graph-Based Route Optimization
A Java-based routing engine that computes fuel-efficient travel paths using classic graph algorithms with a focus on performance and maintainability.
Python · Graph Algorithms . Traversals
AP TRANSCO — Data Science Engineer Intern
Contributed to data science and engineering initiatives for power distribution operations — building ML workflows, automating analytical pipelines, and supporting data‑driven decision making.
IEEE Conference
An Empirical Comparative Study on Pruning and Quantization Algorithms for Model Compression
A comparative study of neural network compression techniques, evaluating pruning and quantization approaches to improve inference efficiency while preserving predictive performance.
Build for clarity before cleverness.
Measure before optimizing.
Keep systems modular.
Automate repetitive work.
Never stop learning.