Software Engineer · Document AI & Data Pipelines · Cybersecurity Research
I build pipelines that turn scanned documents into data people can trust: OCR at scale, vector search, and LLM workflows designed to fail loudly instead of silently. I also find access control flaws in production systems and report them privately.
31,887
55,106
140K+
3
48 h
1
Document AI, data pipelines and ML infrastructure I built and run in production, with the bugs, rejected designs and measurements included.
Recovering the printed page number for 83,807 textbook sections after OCR had discarded page positions and the PDFs turned out to have no text layer.
Semantic duplicate detection and exam forecasting over a ~40,000-question WAEC archive.
A six-stage pipeline turning two decades of WAEC and JAMB papers, from printed booklets to screenshots to inconsistent Word documents, into a classified, queryable corpus.
An event-driven GPU pipeline turning 102 scanned curriculum textbooks into clean, structured Markdown, unattended, recovering from the failure modes of open-source ML inference at scale.
blordgroup.ng
The group’s official website, which I built and deployed as sole developer. It runs on Next.js behind Nginx on an Ubuntu server, and it was built for phones first because most visitors arrive on slow mobile connections.
Each finding went to the organisation that owned the risk, or to the national CERT for a government system, before anyone else. Summaries only: no endpoints, payloads or reproduction steps.
Broken access control (IDOR)
A document retrieval path in the Commission's document management system returned incorporation records and director identity documents without checking who was asking. I reported it through ngCERT, which validated it within 48 hours.
Broken access control
A flaw on the university's learning platform exposed other students' personal data. I reported it directly to the CTO. The university fixed it and paid a bug bounty.
Broken access control
While using the app as a customer, I found that an account lookup returned any user's phone number and wallet balance from their email address alone. I confirmed it on my own account only and reported it privately.
I am open to engineering roles and contract work in document AI, data pipelines and web.
Vulnerability reports go to the security address.