I'm a software engineer with experience building and shipping AI agents and generative AI pipelines in production. I'm interested in how agentic systems fail, and how to constrain them.
Agent pipelines, drawn out
Two of my projects drawn as pipelines. Pick one and watch a request move through it, then select a box to see what it does.
Request trace
Projects
Crashout
A data analysis app. Upload a dataset, ask questions in plain language and get answers with interactive charts. React and TypeScript on the front, a FastAPI backend, and a LangGraph agent on Claude that runs its code in an E2B sandbox.
Personal projectRedZone
Monitors conflicts in the ACLED data and builds a dossier for each one from trusted sources. Casualty figures are checked against those sources, answers draw on the dossier, and the site links people to humanitarian aid.
In progressAgentic AI safety
Ongoing reading and study of how agentic systems fail and how to keep them inside their limits.
OngoingWhat I build with
Frontend
- React
- TypeScript
- Vite
- Tailwind CSS
- Recharts
Backend
- FastAPI
- DuckDB
- Pandas
- Supabase
AI
- Claude
- LangGraph agents
- Tool calling
- E2B sandboxes
- Generative AI pipelines
Between a demo and production
I'm Quadri, a software engineer with a BSc in Computer Science from the University of Lagos. I build AI agents and generative AI pipelines for production use.
What keeps pulling me in is the gap between an agent that works in a demo and one you can trust: what it is allowed to touch, where a person checks its work, and how you notice when it goes wrong. It's the question I keep coming back to.
Let's talk.
Questions, ideas or feedback on any of this are welcome. Email is the fastest way to reach me.
abdulmalikquadri007@gmail.com