B.Tech CSE · MNNIT Allahabad
I’m Nishant — I build across threat intelligence, secure systems, and machine learning, with a focus on getting the details right.
↓A FEW THINGS I’VE PUT INTO THE WORLD ↘
A mix of deep systems, curious experiments, and software made to be used.
An AI-powered news platform with a scalable ingestion pipeline, semantic retrieval, and conversational search. Multi-agent processing extracts and categorizes stories; hybrid recommendations rank them using reader feedback and freshness.
A kernel-native stateful firewall with a four-stage XDP pipeline, in-kernel TCP connection tracking, and firewallctl for live rules, connection inspection, and telemetry.
AlphaZero-style self-play with policy-value networks and asynchronous MCTS. PatchNet cuts model parameters by ~60% across board sizes; the agent reached >95% wins over alpha-beta baselines.
A real-time audio streaming and transcription system: ESP32 microphone audio over WebSockets, a Node.js and Redis coordination server, and a Java desktop client for live waveforms and AssemblyAI transcription.
A secure document management platform built for the Indian Air Force’s private captive mobile network, with end-to-end encryption, replay protection, role-based access, and configurable document security.
An offline-first Android planetarium built on Stellarium Web Engine. Gyroscope sky tracking, an AR camera star-map overlay, and a bundled catalogue of ~50K stars and deep-sky objects.
↳ Most of the fun stuff lives in the details. Every project has a rabbit hole — pick one and say hi.
LEARNING AT REAL-WORLD SCALE ↘
A couple of teams, real constraints, and software built for people who depend on it.
DON’T JUST TAKE MY WORD FOR IT ↘
A small window into a few things running out in the world. Pick a project and have a look around.
Some sites don’t allow themselves to be embedded. The direct link is always there if the window stays blank.
A LITTLE CONTEXT ↘
I’m pursuing a B.Tech in Computer Science & Engineering at MNNIT Allahabad (CPI 8.67). I like going deep — from kernel networking and threat intelligence to how an agent learns a game — then turning that understanding into software people can rely on.
Outside of code: competitive programming, chess, flute, and the occasional astronomy rabbit hole.
Have something interesting in mind?Write me an email ↗Agents, reinforcement learning, and the engineering to make them useful outside a demo.
✳C / C++ · Python · Java · TypeScript
React · Next.js · FastAPI · LibTorch · Socket.IO
Linux · Docker · eBPF / XDP · PostgreSQL · Redis · AWS
A FEW MILESTONES ALONG THE WAY ↘
GOOD IDEAS ARE BETTER SHARED ↘
25.4920° NOpen to interesting ideas, collaborations, and conversations about things worth building.
nishant040305@gmail.com↗