Work Experience
Freelancer
Took on the challenge of rebuilding a UK client's AI video generation platform from the ground up.
- Architected and delivered a production-ready AI video platform from ground up, scaling client revenue from $0 to $100k+ monthly ARR through pipeline stabilization and UX redesign.
- Owned end-to-end development across Frontend, Backend, Auth, Payments, GenAI, and AWS infrastructure. Reduced monthly refunds by 100% ($4.5k to $0) and improved customer retention.
ProofofSkill
Joined a fast moving team focused on helping companies validate skills through real world assessments.
- Accelerated client delivery by shipping 50+ production PRs across 5 applications using Next.js, TailwindCSS, and AWS, reducing feature deployment time by 40%.
- Conceptualized and built an AI chat agent for recruiters using RAG architecture, improving candidate matching efficiency and reducing screening time by 60%.
- Collaborated with design team to deliver pixel-perfect UIs and high-converting landing pages. Eliminated 100% of IVS recording failures, ensuring zero interview disruptions for 500+ monthly assessments.
Freelance Software Engineer
Took on independent projects that pushed me to solve real world problems with AI and web technologies.
- Engineered an AI productivity app processing 10k+ emails via Gmail API with RAG and pgvector, enabling natural language queries with 95% accuracy for personalized insights.
- Delivered mission-critical ID generation platform for 1,500+ government staff under 3-week deadline for FIFA International event, achieving 100% on-time delivery with zero downtime.
Skills
Check out my latest work
I've worked on a variety of projects, from simple websites to complex web applications in Gen AI. Here are a few of my favorites. You can find my resume here
ChatUI
With so many great LLMs out there, I got tired of juggling subscriptions, so I built ChatUI. Bring your own API keys for any OpenAI API compatible model, and chat away with a mobile friendly PWA interface. I also added support for up to 5 team members to brainstorm together with an LLM. This project taught me a ton about optimizations and best practices, and I'm now evolving it into a Microservices architecture with a Go backend and Python for LLM logic. Over 140 users and counting!
StockGPT
I wanted a smarter way to analyze stocks, so I built StockGPT, an intelligent terminal based tool that uses LLMs with tool calling to fetch real time financial data via yfinance. Powered by LangChain and Gemini 2.0 flash, you can ask natural language questions and get insights across 125 different stock related values. I also added plotly graph objects to visualize and compare multiple stocks over time in interactive plots.
Twitter Clone
I wanted to sharpen my UI development skills and explore GraphQL as a different way to talk to backends, so I built a Twitter clone. Used Next.js for the frontend, Node.js for the backend, and Prisma ORM for PostgreSQL. Added Redis for caching and rate limiting, plus Google OAuth for auth. Deployed it on AWS in a scalable setup. (The instance might be offline since there aren't active users, but the code's all there!)
PDFQuery
Ever wished you could just ask questions about a PDF and get answers? That's PDFQuery. Built with RAG, it uses LangChain for language processing and FAISS for similarity search. The interface is a simple Streamlit app powered by Python, and it runs on the gemini-flash-2.0 model via the Gemini API.
Get in Touch
Want to chat? Just reach out to me here and I'll get back to you as soon as I can!
You scrolled so far!
Book a Call