I research EU transport economics at the International Road Transport Union (IRU) in Geneva, building econometric models that forecast freight demand across the 27 EU member states. My academic work applies causal inference: my bachelor’s thesis used the synthetic control method to measure the GDP impact of Western sanctions on Russia. Alongside this I build AI tools and multi-agent systems, and I direct and verify their output.
Experience
Junior Research Econometrician
International Road Transport Union (IRU), Geneva, Switzerland
Nov 2025 – present
Build econometric models that forecast freight transport demand across all 27 EU member states.
Deliver short-term and long-term transport demand projections used in IRU policy briefs.
Process and analyse EU transport datasets in Python and R.
Econometric Intern
International Road Transport Union (IRU), Geneva, Switzerland
Sep 2023 – Oct 2025
Developed models to project short-term and long-term transport demand.
Carried out market analysis to support business expansion.
Extracted, cleaned and analysed EU transport industry data.
Assistant to WTO Counsellor
Permanent Mission of Uzbekistan to the UN Office, Geneva, Switzerland
Jan 2023 – Mar 2023
Prepared written reports for the UN Office and international organisations.
Produced daily reports on socio-economic conditions across global regions.
Liaised with embassies and supported colleagues in WTO policy work.
Selected work
Each entry shows its status and, where a check was run, what it covered.
SPIE hypothetical buyout (LBO)
Verified
A hypothetical buyout of SPIE at a 20% takeover premium and a 12x EBITDA exit loses sponsor money: EUR 9.24bn of sponsor equity returns EUR 7.41bn, which is 0.80x money back or −4.3% a year (hypothetical, reference date 31 Dec 2025); without the premium it breaks even.
Checked: the model was independently rebuilt from its written contract and matched to floating-point precision (largest gap 2e-12); the workbook was recalculated in a separate engine.
Limit: a hypothetical scenario on analyst assumptions (exit multiple, financing terms, tax); the break-even exit is about 14.0x. Not a transaction, not investment advice.
SPIE cash and valuation (DCF)
Verified
To justify SPIE’s market reference of EUR 9.99bn, margins would need to rise from 6.7% to about 8.1%; the conditional operating value is EUR 8.1bn (hypothetical reconstruction), and FY2025 cash after leases and investment was about EUR 554m.
Checked: every input was traced to SPIE’s annual reports (page-cited); the valuation was rebuilt independently from the stated assumptions and matched to about 1e-11.
Limit: a reconstruction after the fact, using reports published in 2026; terminal value is 75% of the total. Not a price target.
Independent audit of an AI-built project
Partly verified
Of 91 claims checked, 68 were supported, 15 partly and 1 contradicted, and the rest were unverifiable, untested or historical; the audit found the buyout’s loss was essentially the 20% premium (hypothetical), which the original headline did not say.
Checked: a second AI, commissioned and directed by me, checked a first AI’s research line by line, with independent rebuilds of every model, workbooks recalculated and sources re-fetched.
Limit: 129 supporting documents were scanned rather than read line by line, as the report states.
EU macro data pilot
Verified
Of 28 comparison points against a free data aggregator, 6 matched, 18 were missing and 4 differed, so the work went to seven official series from Eurostat, DG ECFIN and the ECB.
Checked: all comparisons were reproduced from the raw downloads, and the official values were re-fetched one day later, with no revisions.
Limit: a small sample of series; not a reliability verdict on any provider.
Bachelor’s thesis: economic impact of sanctions on Russia
Graded 5.75/6
Applied the synthetic control method to measure a 5–6.1% decline in Russia’s GDP per capita from Western sanctions after 2014.
Study tool for the CFA® Program
In progressNot independently checked
Turns the 10 CFA® Program curriculum volumes into layered study notes with AI.
Education
Bachelor of Arts in Economics
University of St. Gallen (HSG), St. Gallen, Switzerland
2020 – 2024
GPA 4.9/6.
Bachelor’s thesis, 5.75/6: “Assessing the Economic Impact of Sanctions on Russia”.
International Baccalaureate
College Champittet, Pully, Switzerland
2015 – 2020
Grade 34/45.
Level I candidate in the CFA® Program (exam February 2027)
Skills and languages
Technical
Python, R, JavaScript/Node.js, Claude API, LLM orchestration, multi-agent systems, data pipelines.
Uzbek and Russian (native), English (C1), French (B1–B2).
How I work with AI
Direct
I ran multi-agent research with separate producer and reviewer agents, under written contracts.
Verify
I challenged the first portfolio as too shallow and commissioned an independent audit by a different AI; it found the buyout headline hid its main driver.
Disclose
For every project I state what was checked and what was not.
Worked example
In the independent audit above, a second AI checked a first AI’s research line by line, and I commissioned and directed that check. Every model was rebuilt independently, the workbooks were recalculated and the sources were re-fetched. 129 supporting documents were scanned rather than read line by line, and the report says so.
Contact
Interested in working together or have a question? Reach out directly.