I'm a Physics undergraduate at the University of Rajshahi, Bangladesh, working at the intersection of physics, mathematics, and computing.
Rather than learning technologies only through tutorials, I prefer building things from the ground up - from numerical solvers and multi-node computing clusters to production web applications and self-hosted infrastructure.
Currently exploring:
Scientific computing & numerical methods
High-performance and parallel computing
Systems programming
Full-stack application architecture
Self-hosted infrastructure & DevOps
Machine learning
Mathematical modeling and computational physics
For more go to this section.
My physics background naturally pulled me toward computational methods.
Scientific Computing
Numerical methods for mathematical and physical problems - with a particular interest in ODE systems, numerical integration, optimization, and computational modeling.
High-Performance Computing
Experimenting with building and managing a small multi-node HPC environment, with a long-term goal of scaling computational physics problems across multiple CPU cores and machines.
A programming language ecosystem featuring a tree-walk interpreter built from scratch in C++, a native code formatter (vspfmt), and a dedicated VS Code extension with TypeScript-style syntax highlighting.
- Lexical Analysis & Parsing - Tokenizer and recursive-descent parser generating clean Abstract Syntax Trees (AST).
- Runtime & Execution - C++ tree-walk interpreter evaluating expressions and handling dynamic variable scoping.
- VS Code Integration - Custom extension providing syntax highlighting, auto-formatting via
vspfmt, and language tooling. - Documentation & CI/CD - Automated site deployment via GitHub Pages and automated cross-platform builds.
Focus: language design · AST evaluation · VS Code extension development · C++ system programming
🌐 Documentation · 🖥️ Source · 🧩 VS Code Extension for language support
A production-oriented, self-hosted management platform for diagnostic centers.
- Patient and invoice management
- Diagnostic test management
- Doctor commissions and discounts
- Payment and due tracking
- Sales and financial reporting
- Audit logging & role-based administration
- Printable invoices and reports
Also involved database migrations, deployment, reverse proxies, Windows packaging, backups, and self-hosted infrastructure - making it as much a systems project as a web application.
An anonymous peer-to-peer chat experience built around a playful token economy - match with a random fox, chat for a while, earn berries, and come back for more.
- Instant Matching - find and connect with a random anonymous fox
- Berry Economy - earn berries by chatting, lose them when skipping
- Timed Chats - 2-minute initial sessions with 5-minute extensions
- Real-time Messaging - instant messages, typing indicators, and reactions
- Persistent Sessions - rejoin a conversation after an accidental disconnect
- Responsive UI - designed for mobile, tablet, and desktop
- Anonymous by Design - no accounts, no unnecessary tracking
Focus: real-time systems · WebSockets · anonymous matching · state management
A small personal cluster built to experiment with parallel scientific workloads - understanding not just how to write numerical programs, but how computation behaves when workloads are distributed across CPUs and machines.For more projects visit my repositories.
Languages
Web & Application Development
Databases & Backend
Scientific Computing & Mathematics
Systems & CLI
Infrastructure & DevOps
Version Control & Development Tools
Scientific Computing
├── Numerical ODE Solvers
├── Parallel Computing
├── MPI / Distributed Systems
└── Computational Physics
Systems Programming
├── Rust
├── Linux
└── Performance Engineering
Production Software
├── Next.js
├── PostgreSQL
├── Docker
└── Self-hosted Infrastructure
I'm especially drawn to the overlap of these areas - using low-level, high-performance computing ideas to solve real scientific and engineering problems.
Learn the theory. Build the system. Measure the result.
I like understanding what's happening underneath the abstraction - whether that's a numerical algorithm, a database query, a network request, or a CPU instruction.


