I’m Usama, a self-taught developer building web apps, Android experiences and AI tools. Curious about the interface. Careful about what happens underneath.
The person behind the work ↗A little context
I’m a self-taught Software & AI Application Developer. I build interfaces, connect them to APIs, and debug the path from a user action to stored data or a model response. I learned by building things I wanted to use and testing where they broke.
One of those ideas was a fitness app I wanted for myself. Working on FitDesi encouraged me to build more apps, websites and workflows. Today I work across Android, web development and backend APIs, with an interest in making AI useful inside real software.
Read my story ↗FastAPI, React, Kotlin, Jetpack Compose, Apps Script and model APIs.
Python, JavaScript, TypeScript and Make. Exploring LangGraph, retrieval, Docker and AI security.
Independent projects / selected work
Android · Product design
A personal fitness companion with structured planning, live sets and training history.
Engineering decision
The BASIC experience uses local FitDesi knowledge and structured workout logic. Real Android captures show the interface; the app remains private and pre-launch.
Kotlin / Compose / Local-firstAI application · React / FastAPI
An AI command center that makes agent handoffs visible.
Engineering decision
Planner, Research, Executor and Reviewer stages are paired with an audit trace. The hackathon prototype supports model and mock execution modes.
React / FastAPI / Agent workflows
Web design · Technical writing
A custom publication for practical development and AI guides.
Engineering decision
A real Blogger theme and published article, presented at desktop and mobile widths. These are authentic captures, not a redesigned product concept.
Blogger / Theme engineering / WritingThe path so far
From wondering how things work
to building things of my own.
I’ve been a tech enthusiast for as long as I can remember. I knew a little coding, but there was a gap between the ideas in my head and what I could build. Curiosity kept pulling me back.
Discovering AI changed what felt possible. Tasks I imagined taking weeks could become an experiment in a day. But getting an answer was only the beginning: I had to understand it, change it and find out where it broke.
I started building a fitness app for myself. A simple personal need became FitDesi: planning workouts, tracking sets and making the experience useful without a connection. Each improvement gave me another reason to keep learning.
Explore FitDesi ↗With NexusOps I explored a workflow where planning, research, execution and review could be inspected separately. That made the execution trace as important as the final report.
Explore NexusOps ↗FlyRank gave me a place to practise backend development and receive feedback. With ByteEdition, I worked on explaining technical steps clearly enough for someone else to follow.
Explore ByteEdition ↗I want to make useful applications, keep improving my engineering judgment and work with people who care about how software behaves in practice.
Get in touch ↗How I work
A clear process.
Room for discovery.
Start with the intended behavior. For a broken API, reproduce the failing request and capture the response.
Output · Reproduction and expected behaviorChoose the smallest useful result. FitDesi separates everyday offline fitness tools from later advanced coaching.
Output · Scope and constraintsUse AI to explore options and assist implementation. Review changes against the intended flow and existing code.
Output · Reviewed changeCheck real inputs and outputs. For backend work, inspect validation responses and verify what happens to saved data.
Output · Checks and known limitsDocument setup, decisions and remaining work so another person can use or continue the project.
Output · Setup notes and next stepsWhere I can help
Pakistan · PKT / UTC+5 · Project and career enquiries
