About
I’m Ivan Gorban, a Staff Data Scientist working at the intersection of machine learning, causal inference, experimentation, and applied AI. My work is mostly about turning ambiguous product and business problems into models, metrics, experiments, and decision systems — especially in marketplace environments where prediction alone is rarely enough.
What I work on
Machine learning for products — I work on applied machine learning systems for real product and marketplace decisions: pricing, churn prevention, lifecycle modeling, customer and supply behavior, and operational trade-offs. I’m especially interested in models that do not just predict outcomes, but help teams decide what to do next.
Causal inference & experimentation — I work with experiment design, causal measurement, uplift estimation, switchback tests, regression-based analysis, and practical methods for separating correlation from intervention effects. My focus is on using causal thinking to make product decisions more reliable, not on treating experimentation as a reporting layer after the fact.
Applied AI & LLM systems — I explore practical applications of AI and LLMs in workflows, tooling, education, and knowledge work. This includes agentic coding workflows, AI-assisted analysis, tutoring systems, and the question of how to make AI useful without replacing the parts of thinking that actually build expertise.
Research & writing — I write about causal inference, experimentation, AI, automation, data science practice, and the economic and educational consequences of new technology. My goal is to make technical ideas clear without flattening them into slogans, and to connect methods with the real decisions they are supposed to support.
Background
Careem (an Uber company) — Staff Data Scientist · Dubai, UAE · 2022–present
Careem is a super-app marketplace operating across the Middle East. I joined as a Senior Data Scientist and was promoted to Staff in 2025. My work spans the full arc from early prototyping to production systems: dynamic pricing for the QUIK grocery vertical, a captain churn-prevention system, adaptive ETA prediction, a marketplace simulation environment, and an experimentation platform I’ve operated across data science, engineering, and product management functions. I’ve also built an agentic LLM framework for internal workflows and led a university research collaboration that produced four publications.
MegaFon OJSC — Lead Data Scientist · Russia · 2018–2022
MegaFon is one of Russia’s major telecommunications operators. In the Big Data Department, I led a team of data scientists and business analysts working on strategic and operational problems: retail-network footprint strategy, optimal pricing, marketing-mix modeling, and a reinforcement learning–based Next-Best-Action system with causal nudge planning. The work combined long-horizon strategy (where to open or close retail locations) with shorter-cycle optimization (which offer to show which customer, and why).
Pompeu Fabra University — Researcher, Social Media Group · Barcelona, Spain · 2016–2018
At the Social Media Group I worked on network science and social media analysis. This period laid the groundwork for the graph-theoretic and observational-data intuitions that inform my current causal inference work.
Education
- M.A. Economics (Finance, Data Analysis) — New Economic School, Moscow, 2016
- M.S. equivalent in Mathematics — Moscow Aviation Institute, 2011
Publications
Research I’ve contributed to or led:
- “Restaurant Menu Categorization at Scale: LLM-Guided Hybrid Clustering” — INLG 2025
- “Unraveling the Implications of Silent Labor Time in the Gig Economy” — INSEAD working paper, 2025
- STEF-DHNet — arXiv:2306.14476, 2023
- “Do Recommendation Systems Work” — SSRN 4284103, 2022
Full listing on the Research page.
Projects
Quoin — a developer workflow library that gives developers a structured way to plan, implement, and review work with AI agents without giving up control of the codebase.
OpenClaw-based AI tutor — for a child, in progress.
Talks & demos
Get in touch
The best ways to reach me: