Breaking &
Rebuilding
I enjoy understanding how systems work, breaking them apart, and rebuilding them into something useful.
Curiosity driven. Constantly building.
Hi, I'm Debajeet Mandal, a Computer Science & Data Science student. I enjoy learning, debugging, and engineering systems where data meets automation.
My programming journey started seriously in 2025 by building ATOM, an offline desktop voice assistant, while learning Python. That project sparked a deep interest in machine learning, NLP, and systems design. Since then, I've spent my time solving DSA challenges in C++ on LeetCode, writing automation workflows with n8n, and diving head-first into open-source software, contributing to libraries like Visdom, Astropy, and SymPy.
I believe in continuous improvement—learning through building, finding bugs, seeking code reviews, and writing software that solves actual, real-world problems.
Why do I build?
Curiosity driven.
Not because I chase trends,
but because I am obsessed with understanding how systems work.
What motivates me isn't simply writing software—it's the idea of creating tools that genuinely improve the way people work and think. I've always been inspired by engineers who treat technology as a means to solve difficult problems rather than an end in itself.
That perspective shaped how I approach programming today. I enjoy taking apart complex systems, understanding why they behave the way they do, and rebuilding them with clarity and purpose. Whether I'm debugging a stubborn issue, designing a new feature, or contributing to open source, I'm driven by the belief that thoughtful engineering can turn ambitious ideas into something real, reliable, and genuinely useful.
How I Build
My operating system. I don't just code—I build with systems clarity, continuous automation, and long-term maintainability.
Understand First
I rarely begin by writing code. Before changing a line, I trace data flow and inspect architecture. Good software emerges from deep understanding.
Reading Over Writing
Reading production code has taught me more than tutorials ever could. Every PR, review comment, and bug report reveals how experienced engineers think.
Automate Repetition
If I solve something twice, I look for a way to never solve it manually again. I enjoy removing operational friction more than adding bloat.
Value Maintainability
Elegant software isn't the cleverest algorithm—it's the one another engineer can understand six months later. Build. Review. Refine.
“I don't chase perfection. I chase continuous improvement.”
Still learning.
Still breaking things.
Still building them.
Still chasing elegant systems.
Still contributing. Still curious.
Still Learning
Every day brings a new problem and a new lesson.
Still Building
Turning ideas into systems that solve real problems.
Still Breaking
Breaking things to understand them at a deeper level.
Still Chasing
Chasing elegant systems and meaningful impact.
Still Learning
Every day brings a new problem and a new lesson.
Still Building
Turning ideas into systems that solve real problems.
The best engineers never stop being students.
The future is engineered one system at a time.
Built
not
promised
Friday AI
Voice AI • Local OS Agent
Multimodal voice agent using LiveKit WebRTC and Gemini Realtime API for hands-free local system automation.
Explore More↗
Project Atlas
Developer Tools • AI Agent
Hybrid repository indexer & bug hunter using Groq LLaMA 3.3 and Gemini to plan fixes and draft pull requests.
Explore More↗Bharat Nirman
GIS Analytics • Smart City
Sustainable urban infrastructure GIS planning dashboard displaying multi-layer municipality zoning data.
Explore More↗
S.T.E.M
NLP • Market Sentiment
Stock market sentiment prediction engine crawling Wikipedia summaries and correlating scores with financial trends.
Explore More↗Developer Workstation
Every technology I've learned has been acquired by building real projects—not by collecting certificates. Watch my development environment boot up.
Engineering Journey
This is not a chronological list of achievements. It is the story of how my curiosity became an obsession, how I learned systems programming, and how I started contributing to global scientific software.
Curiosity Sparked
I had always loved technology, but it was a black box. I found myself wondering how software actually worked under the hood. I wanted to understand the mechanics behind the tools I used every day, which made me decide to stop simply consuming technology and start building it.
ATOM
I built ATOM, an offline desktop AI assistant, while I was still learning the fundamentals of Python. It featured custom speech recognition loops, file system management, and script execution shortcuts.
Certifications & Hackathons
Factual history of participation, technical challenges solved, and workshop completions.
GDG Devcation Delhi 2026
Collaborated in designing and building a responsive web prototype under Google-specific mentors during a high-intensity 36-hour sprint.
EY Techathon
Designed AI automation templates and agent flow logic for modern enterprise resource planning systems.
TENET Hack 25
Built decentralized interface concepts and authentication pipelines leveraging Web3 systems.
AI & ML at IIT
Completed comprehensive practical training in neural networks, classifier training, and geospatial data analytics frameworks.
OpenAI Buildathon
Developed and benchmarked system prompts and agent capabilities using the Gemini and OpenAI API suites.
Singularity Tech & Media
Engineered visually responsive dashboard concepts and structured client onboarding databases.
AI Agent Workshop
Hands-on implementation of multi-agent orchestration frameworks (CrewAI, LangGraph) for visual parsing.
GenAI Challenge
Engineered prompt logic chains and JSON schema output mappings using modern LLM APIs.
Common Questions
Debajeet Mandal
Let's build
something meaningful.
Whether it's open source, AI systems, automation, research collaborations, hackathons, or simply discussing ideas—I enjoy building with curious people.