Michael T.
Data Lead
Experience
Data Lead
Gower Street Analytics
Reduced ML deployment time from weeks to days by building robust inference endpoints of various complexity (AWS Lambda, Flask, FastAPI, Dagster).
Increased data processing capacity 100x (200GB to 20TB) by implementing a scalable AWS data lake (Iceberg, Parquet).
Optimized core SaaS PostgreSQL database on AWS RDS (read replicas, schema redesign, query optimization), enabling expansion from 6 to 90 markets.
Cut manual metadata work by 90% through agentic workflows (ChatGPT, Playwright, Python), accelerating data ingestion.
Developed film revenue forecasting and film similarity models (TSLib, scikit-learn, HDBSCAN) that secured 3 major studio partnerships.
Worked closely with the executive team to define the strategic roadmap for data-driven products and shape the future direction of the company’s offerings.
Delivered real-time analytics dashboards (Metabase) reducing ad-hoc requests by 40% and support tickets by 80%.
Designed automated PDF reporting system (Polars, LaTeX), achieving 96% positive feedback and helping win two $500k contract renewals.
Built and led a high-performing data team: led hiring (data scientists, software engineers), onboarding, mentoring, and line management.
Senior Data Scientist
Gower Street Analytics
- Developed and deployed a time series analysis model to predict film ticket prices in multiple countries (USA, UK, Germany, Australia), achieving an average accuracy of 93% over a 12-month period.
- Engineered a gradient-boosted model (pandas, scikit-learn, PostgreSQL) to predict lifetime film earnings from partial data, achieving a mean percentage error of 1% and saving analysts approximately 10% of their weekly time.
- Automated the generation and deployment of weekly reports for both internal and external stakeholders, leveraging a tech stack including AWS Athena, R/KnitR/LaTeX, GitHub, Drone, and Docker Swarm.
- Established and maintained a Docker-based environment for data science projects on remote servers, which promoted team collaboration and enabled rapid experimentation. Mentored colleagues on best practices and usage.
Data Scientist & Mathematical Modeller
Skyscanner
- Developed, tested, and deployed statistical and mathematical models to optimize budget allocation across advertising markets and channels, directly informing monthly decisions on budgets exceeding $6,000,000. These models led to a 9% year-on-year increase in advertising ROI.
- Conducted advanced machine learning work, including revenue attribution modeling, customer lifetime value prediction, causal impact analysis, and ad revenue and cost-per-click optimization.
Teaching Assistant
Heriot-Watt University
Assisted with lectures, tutorials and marking in the following modules: Data Analysis; Probability & Statistics; Financial Mathematics; Differential Equations; Mathematical Modelling.
Languages
Education
Heriot-Watt University
Ph.D. · Mathematical Modelling · Edinburgh, United Kingdom
National Technical University of Athens
M.Sc. · Applied Mathematical and Physical Sciences · Athens, Greece
Similar Freelancers
Discover other experts with similar qualifications and experience
- English
- Deutsch
2025 © FRATCH.IO GmbH. All rights reserved.