Hyprland
CPU 17%
RAM 3.2G
eth0
VOL 64
niranjan
Data Scientist & AI Engineer
I turn messy data into models people actually use — from image classifiers to intent-driven agents.
Data Scientist & AI Engineer
I turn messy data into models people actually use — from image classifiers to intent-driven agents.
First-year B.Tech Artificial Intelligence & Data Science student at KPR Institute of Engineering and Technology (KPRIET), Arasur. Focused on machine learning, computer vision, and building useful things with LLMs. I like small, reliable systems and honest metrics.
A Smart India Hackathon 2026 platform: a citizen SOS app and authority dashboard for locating flood victims, with automatic Bluetooth-mesh fallback when the network is down and live CWC river-gauge risk zones.
A full-stack AI dashboard that runs your inbox: local LLM summarization via Ollama, spam classifier, AI-composed emails, and real-time voice interface.
An outbound calling bot that talks to real people. FastAPI orchestrates Exotel telephony with a Pipecat pipeline — Deepgram STT, OpenAI LLM, Cartesia TTS.
A production marketing site for a real tuition centre. Next.js 15 static export backed by a Google-Sheets CMS.
A Linux server running Docker behind Cloudflare Tunnels and a Tailscale mesh, self-hosted Supabase, a CI runner, Grafana + Loki observability, and an autonomous DevOps agent that runs ops from natural-language commands.
An MCP AI-tool gateway, an LLM-driven recon agent for authorized homelab security testing, and a defensive endpoint guard that rate-limits and bans probes on a public MCP endpoint.
Cortex AI daemon, smart-terminal, Mailing Agent, Password Manager, and more.
Best reached by email. I usually reply within a day.
niranjan.dev@proton.meData Scientist & AI Engineer
I turn messy data into models people actually use — from image classifiers to intent-driven agents.
First-year B.Tech Artificial Intelligence & Data Science student at KPR Institute of Engineering and Technology (KPRIET), Arasur. Focused on machine learning, computer vision, and building useful things with LLMs. I like small, reliable systems and honest metrics.
A citizen SOS app + authority dashboard for locating flood victims, with automatic Bluetooth-mesh fallback and live river-gauge risk zones.
A full-stack AI dashboard that runs your inbox: local-LLM summaries, a spam classifier, AI-composed emails, and a voice interface.
An outbound calling bot that talks to real people — Exotel telephony bridged to a Pipecat voice pipeline over WebSockets.
A local-first personal AI daemon, a natural-language terminal assistant, and more.
Best reached by email. I usually reply within a day.
niranjan.dev@proton.me Download résumé (PDF) ↗Want the full experience?
Designed for laptop/desktop screens.matrixDigital rainsudo hire niranjan🎉neofetchSystem infoplay datasetsML playground