14 y/o from Pakistan. Here is everything about me. Who I Am and How I Work.

Sumama Jameel, a 14-year-old developer in Karachi, Pakistan, has shipped nine products that he designed and built by directing AI to write, test, and debug the code. His primary project, Megabit, is a data generation engine that produces verified training data for AI coding models by having an agent recreate real open-source code changes in a sandbox, and he is preparing sample datasets for AI labs while applying to Y Combinator's Winter 2027 batch. Other released tools include Mirage, an open-source OpenAI-compatible AI gateway built in Rust, and Whelmer, a browser debugging tool for AI agents.

Who I Am My name is Sumama Jameel. I am 14 years old. I live in Karachi, Pakistan. I study at Tameer-e-millat model school, in Karachi. I am a product architect and builder. I design systems, make plans, research market, define architectures, set constraints, and direct AI to build the software. I have shipped nine products this way. Every system I have built exists because I understood the problem, designed the solution, and executed it through the most efficient way possible. Email: sumamajamil2005@gmail.com mailto:sumamajamil2005@gmail.com GitHub: github.com/Sumama-Jameel Location: Karachi, Pakistan What I Build I build tools that solve real problems. Here is what I have shipped: Megabit — A data generation engine for AI coding models. It takes real code changes from open-source repos, hides the solution, and makes an AI agent recreate the change in a real sandbox. Every command is executed. Every error is real. The output is verified training data for AI labs. Mirage — An open-source AI gateway. It gives you a single OpenAI-compatible endpoint that routes to ChatGPT, DeepSeek, Gemini, and other models using free internal APIs. Zero API keys. Zero per-token billing. Built in Rust with a custom headless browser engine. Grasper — A reading companion for hard science books. It shows you one page at a time, explains it from first principles, asks you five questions, and only unlocks the next page when you prove you understand. Supports PDF, EPUB, and plain text. Validator — An agent skill that validates a product idea before you build it. It researches the market, finds every competitor, destroys your assumptions, and gives you a GO, CONDITIONAL GO, or NO GO verdict backed by sourced evidence. Whelmer — A browser debugging tool that gives AI agents full visibility into what a browser is doing. Every DOM change, every network request, every console message. Nothing hidden, nothing filtered. Built in Rust. Supports Chrome, Brave, Edge, Opera, and Firefox. SubZero — A privacy-first subscription tracker. It finds hidden recurring charges in bank statements without connecting to your bank. Nine products 6 listed here . All designed by me. All built through AI engineering. Most of them are open source. How I Build I am the architect, define the problem, design the system and set the constraints. I write the specifications, tech stack and user experience flow. I direct AI to write the code, test it, debug it, and ship it. I understand systems at a level that lets me break any problem into clear, executable steps. I know what good architecture looks like. I know what bad architecture looks like. I know when AI produces correct code and when it produces garbage. I review every output. I reject what does not work. I iterate until it does. My workflow: This method lets me move fast. One person with clear thinking and AI as the engineering team can ship what used to require a team of five. My Philosophy I build things that are honest. No fake data. No hallucinated outputs. No marketing fluff. If a product claims to do something, it actually does it. If it cannot do something, I say so. I build things that solve problems I actually have. I do not build for hype. I build because something is broken and I want it fixed. I build in the open. Everything I make is on GitHub. Anyone can read the code, run it, break it, and tell me it is wrong. I want that. I want people to find the flaws so I can fix them. What I Am Working On Now Megabit is my primary focus. I am preparing sample datasets to send to AI labs. I am applying to Y Combinator Winter 2027. I am building the sales pipeline with my business co-founder. The goal is clear: become the standard supplier of verified coding trajectory data for AI companies. The market is 4 billion dollars in 2026 and growing 30 percent per year. AI labs need grounded data. I build grounded data. Connect With Me If you are building AI products, working on training data, or interested in any of my open-source projects, reach out. I respond to every message. No gatekeeping. No ego. Just build.