Mihnea Todosie
I'm a data analyst/scientist with 8 years of experience across data wrangling, predictive analytics, and machine learning.
About
I’m a data analyst/scientist with 8 years of experience across data wrangling, predictive analytics, and machine learning. I specialize in R with strong Python and SQL skills, building scalable ETL pipelines, automating reporting, and designing predictive pricing models that improve ROI and efficiency.\n\nI have hands-on experience with A/B testing, statistical analysis, and data visualization to support data-driven decisions. I thrive on turning complex data into actionable insights and scalable data engineering solutions.
- 7 years' experience
- Joined December 2025
Languages
- EnglishFluent
- Romanian, Moldavian, MoldovanFluent
- Spanish; CastilianIntermediate
- ItalianIntermediate
- GermanIntermediate
- FrenchIntermediate
- JapaneseBeginner
Industry experience
- Software & Internet
- Financial Services
- Gaming
- Professional Services
- Media & Entertainment
Work experience· 4 roles
Built a Python-based OCR invoice pipeline automating data extraction and validation, reducing monthly reporting time by 95%+ (from ~40 h to ~2–3 h). Developed automated financial and management reporting workflows that eliminated manual consolidation across entities. Led the Microsoft Business Central ERP integration, improving accounting transparency and aligning budgeting with realised performance.
Refined margin-curve transparency and pricing in R and Excel and tested provider-specific pricing strategies to optimise portfolio profitability, whilst properly documenting in Jira/Confluence. Devised a partial cashout algorithm for the book, as well as integrating the whole pricing model into the new Back Office and in Tableau. Designed and optimised multi-source ETL pipelines for odds validation and trader operations, reducing latency and manual reconciliation by > 70%.
Developed a fully automated competitor-price ETL pipeline in R (AP + scraping) running every 10 min; scaled runtime from >30 min to <3 min with <1% error. Built match-fixing detection models (“guilty-by-association”) to flag suspicious betting accounts, replacing manual reviews and improving risk management.
Education
Hire Mihnea Todosie today
Post a job and invite Mihnea, or contact them directly. No hidden commission.