Product, AI, risk, and institutions. I’m interested in systems that make complex problems easier to understand, verify, and act on — most of what I’ve built comes back to one question: is this claim true, and can you prove it later.

Systems

Since May 2026 I’ve been building Summativ, a digital assessment platform for Azerbaijani educators. It folds KSQ/BSQ evaluations, student data, and grading into one system, so exam access and result analytics stop living in scattered spreadsheets and group chats.

Before that, in January 2026, I started Verifex alone: a probabilistic identity-match engine now screening 4.8M+ compliance and watchlist records against global sanctions and PEP databases. The hard part isn’t matching exact strings — it’s knowing that “Sədat Nəzərli,” “Садат Назарли,” and “Sadat Nazarli” are the same person across three alphabets and a typo. I wrote a semi-naive log-likelihood ratio framework for that phonetic and transliteration matching, ran it through adversarial benchmarks until it hit 100% F1 with zero false negatives, and built an “Evidence Capsule” schema — SHA-256 hash chaining so every decision stays provable after the fact.

Writing

The other project doesn’t involve any of that. Quaer is an Azerbaijani intellectual media platform I founded in December 2025 — essays, philosophers, concepts, and debates, built into a structured reading library rather than a feed. I own the editorial direction, the visual identity, and most of the writing itself: philosophy, society, and the motives hidden behind how people think and live.

Earlier — Oct 2024 to Jul 2026, remote, as a contract Software Engineer at CİB Biznes — I engineered a custom KYC/CRM architecture integrating third-party compliance APIs for merchant onboarding, wrote the API documentation that streamlined how other developers worked with it, and built Telegram bots that automated tasks and cut manual data entry. Mostly TypeScript and Node.js.

Credentials