Projects

A production-grade blog platform deployed on AWS EC2 (t2.micro) with fully automated CI/CD pipeline. Built with AI-assistance to implement cloud infrastructure, system administration, and DevOps best practices. Features zero-downtime deployments via GitHub Actions, Nginx reverse proxy, PM2 process management, and SSL/TLS encryption with Let's Encrypt. Overcame memory constraints on free tier by implementing swap space and optimizing Node.js build process. Achieved 99.9% uptime with 1-2 minute deployment cycles from git push to production.

An AI-powered news aggregation platform that leverages local LLMs (Ollama with 20B parameter gpt-oss model) to transform complex Guardian articles about India and Pakistan into accessible, simplified summaries. Built with AI-assistance to demonstrate practical application of open-source AI models for natural language processing. Features automated content fetching, intelligent filtering, prompt engineering for content transformation, and AWS S3 + CloudFront deployment with GitHub Actions CI/CD.

A production-deployed AI-powered platform that transforms complex news articles into accessible summaries using local large language models. Built with AI-assistance to showcase practical LLM integration with Next.js. Features the same 20B parameter Ollama model as padho.net, demonstrating content transformation pipeline from raw articles to static MDX files. Deployed to Vercel with automatic CI/CD, proving the viability of local AI models for production content generation.