*
here are some of the projects i've worked on. i love building tools that make developers' lives easier and exploring new technologies along the way.
yt-transcriber
creator and maintainer
open-source cli tool for transcribing and summarising YouTube videos using AI
achievements
- built end-to-end transcription pipeline using Whisper
- designed CLI workflows for multi-language summarisation
- wrote detailed documentation and usage guides
technologies
ai-forum
creator and maintainer
open-source AI-powered discussion platform (in progress)
achievements
- architecting full-stack real-time discussion system
- integrating AI features for content generation and moderation
- designing scalable backend using modern web stack
technologies
protor
creator and maintainer
open-source cli tool for async web scraping and AI-powered site analysis using local LLMs
achievements
- built async scraping pipeline with aiohttp supporting concurrent multi-site crawling
- integrated ollama for local LLM inference with focus modes: technical, seo, and content
- published to PyPI with built-in update checker and rich CLI output
technologies
dripl
creator and maintainer
real-time collaborative whiteboard with hand-drawn rendering, live cursors, and shareable links
achievements
- built three-server monorepo architecture with Next.js frontend, Express REST API, and WebSocket server
- implemented full canvas toolset including shapes, freedraw, frames, and 100-step undo/redo
- engineered real-time collaboration with live cursor presence and scene sync across clients
technologies
sastram
creator and maintainer
AI-powered community forum where questions get resolved through AI synthesis and community validation
achievements
- built modular full-stack architecture with server actions always returning { data, error } and never throwing
- bridged BullMQ job queues to WebSocket clients via Redis pub/sub for real-time AI answer delivery
- integrated multi-model retrieval with Gemini, Exa, and Tavily for inline AI responses via @ai trigger
technologies
meetour
creator and maintainer
self-hostable WebRTC video conferencing platform with waiting rooms, recording pipeline, and Redis-backed signaling
achievements
- implemented full WebRTC mesh with waiting room controls, room lock, passcodes, and admit/reject flows
- built client-side recording pipeline with canvas composite and optional FFmpeg backend merge
- architected Redis-backed real-time signaling fanout with Drizzle ORM for persistent room state