I'm Nandan K S a passionate Full-Stack Developer and MCA student at CMR Institute of Technology, Bengaluru.
I build scalable web applications and AI-powered solutions that solve real-world problems from real-time video conferencing platforms with on-the-fly meeting summaries to fully offline voice assistants that keep your data private.
I'm captivated by the intersection of real-time systems, AI / NLP, and clean software architecture. For me, writing great software isn't just about making things work it's about making them work beautifully.
"Let's build something impactful together."
- 🎓 MCA Student CMR Institute of Technology, Bengaluru
- 🔭 Currently building AI-integrated full-stack applications
- 🌱 Exploring real-time systems, WebRTC, NLP & AI integration
- 💡 Passionate about privacy-first tech, developer tools, and elegant UX
- 🤝 Open to collaborations, new opportunities, and interesting ideas
Languages
Frontend
Backend
Databases
AI & NLP
Tools & DevOps
An intelligent pipeline that transforms any YouTube video into AI-generated summaries, mind maps, and quizzes using Google Gemini + Flask.
| Feature | Detail |
|---|---|
| 📝 AI Summary | Adjustable length (10–50%), cached with SQLite TTL |
| 🗺️ Mind Map | Mermaid-format visual maps generated from transcript |
| 📋 Quiz | Auto-generated Q&A from video content |
| 🔌 Chrome Extension | In-browser video analysis without leaving the tab |
| ⚡ Real-time Updates | Socket.IO processing progress indicators |
Python · Flask · Google Gemini API · yt-dlp · Socket.IO · SQLite
A full-featured video conferencing platform with real-time AI transcription, smart meeting summaries, and ML-based presence detection.
| Feature | Detail |
|---|---|
| 📹 HD Video | WebRTC peer-to-peer video & audio with screen sharing, grid/speaker views |
| 🤖 AI Summaries | Auto-generated meeting summaries & action items via Google Gemini |
| 👤 Face Detection | MediaPipe ML — auto-away detection + missed conversation catch-up |
| 💬 Live Chat | Real-time messaging with typing indicators and emoji support |
| 🗄️ Persistence | MongoDB meeting & chat storage, full REST API + WebSocket events |
React 18 · Node.js · WebRTC · Socket.IO · MongoDB · Google Gemini AI · MediaPipe
Broadcast audio instantly over the web. One device streams; everyone listens through their own headphones — zero Bluetooth, zero extra hardware.
| Feature | Detail |
|---|---|
| 📡 Audio Capture | Browser Tab Audio (getDisplayMedia) or Microphone (getUserMedia) |
| 🔗 Instant Sharing | Unique 7-character room code + QR code generated on demand |
| 🎛️ Host Controls | Pause, resume, stop broadcast, remove individual listeners |
| 🔒 Security | Helmet headers, CORS restriction, rate limiting, stale-room cleanup |
| 🌐 P2P Transport | WebRTC direct audio — no audio data ever touches the server |
React · Vite · TailwindCSS · Node.js · Socket.IO · WebRTC · Web Audio API
A 100% offline, privacy-first Windows voice assistant powered by OpenAI Whisper and a 4-layer NLP pipeline. Your voice, your control, your data.
| Feature | Detail |
|---|---|
| 🎤 Offline STT | OpenAI Whisper or Vosk — no internet, no cloud, no data collection |
| 🧠 4-Layer NLP | Exact → Fuzzy → Phonetic → Semantic matching with weighted scoring |
| 🛡️ Safety | 4-gate validation + command sandboxing (whitelist-only execution) |
| 🖥️ System Control | Open/close any app, smart window & tab control, 40+ keyboard commands |
| ⚡ Performance | Threaded execution, interruptible TTS, startup prewarming, accent support |
Python · OpenAI Whisper · PyAutoGUI · pyttsx3 · NLP Pipelines · Windows OS
Note
The public streak card service can lag because it uses GitHub contribution data, which is reported in UTC and can be cached before it refreshes. For the most accurate and fastest updates, self-host the streak card and disable cache in your deployment.
⭐ Thanks for stopping by! Feel free to browse my repos, star something useful, or reach out if you'd like to collaborate on something impactful.