Statistician and data analyst, working on fiscal analysis and risk assessment.
I build statistical and econometric models in Python and R, and I have shipped production software with Next.js, FastAPI and PostgreSQL. MSc in Statistical Methods and Applications from Sapienza University of Rome. Based in Baku, Azerbaijan.
Grounded in my background - projects, stack, and the statistics behind MyMCMC.
I hold an MSc in Statistical Methods and Applications from Sapienza University of Rome (90/110), with a thesis on adaptive MCMC algorithms, a BBA in Economics from the Academy of Public Administration under the President of the Republic of Azerbaijan (95.8/100), and I'm a SABAH.academy alumnus. That background sits underneath everything I build.
I work as an Adviser in the Fiscal Risk Assessment and Strategic Analysis Sector of the State Enterprises Monitoring Agency under the Ministry of Finance of the Republic of Azerbaijan. Before that I built production systems as a full-stack engineer, with Next.js and TypeScript on the product side and Python, R and SQL on the data side. Most of my tools still live in the terminal.
Civil servant in the Fiscal Analysis and Monitoring Department.
Owned the core architecture of the student recruitment platform, end to end.
Built the center's digital management system and its IT infrastructure.
Built the data architecture for a B2B hiring platform, integrating AI- and NLP-based analysis models.
Product analytics and data science for a real-time artisan booking app.
Sales performance analysis.
Economic analysis internship.
Student recruitment platform connecting prospective students with international universities.
AI-powered B2B hiring platform. Its pricing model is based on active job vacancies and the total number of applicants processed.
Digital management system and IT infrastructure for a vocational education institute.
Cross-platform app connecting clients with local artisans in real time, with live geolocation matching.
My master's thesis research: a from-scratch R implementation of the classical and adaptive Markov chain Monte Carlo algorithms, benchmarked against real Bayesian posteriors from posteriordb rather than synthetic targets - three algorithms, seven posteriors, four chains of 100,000 iterations each.
| Algorithm | Mean ESS | Worst R-hat | Unconverged | Mean error (ref SDs) |
|---|---|---|---|---|
| AM | 9,887 | 1.005 | 0 | 0.022 |
| RAM | 9,086 | 1.005 | 0 | 0.023 |
| RWM baseline | 4,238 | 2.400 | 2 | 0.131 |
Working notes on the engineering behind statistical systems: what the architecture had to look like, what broke first, and what the numbers said afterwards.
Five domains, twenty-four tools, mapped across a sphere.
Sapienza University of Rome
Thesis: Exploration of Adaptive Algorithms
Academy of Public Administration under the President of the Republic of Azerbaijan
Thesis: Tax Evasion, Tax Audit and Analysis
Alumnus
A selection of the coursework and programmes completed outside the degrees. Each certificate is the original document, and the ones whose issuer publishes a verification page link straight to it.
For questions about statistics, data analysis, my research or published notes, write to me by email.
Download Résumé ↓
