Python / Machine Learning Developer — Sports Prediction & Automation

I have an existing Python-based Serie A soccer prediction and betting-analysis model. I am looking for a developer to help audit, improve, automate, and maintain the system. The model is already built; this is primarily improving the existing codebase rather than starting from scratch. Main Responsibilities Audit and clean up the existing Python codebase. Fix the current hyperparameter optimization/tuning system. Standardize Brier score, log loss, calibration, and other performance metrics. Build proper walk-forward/time-series validation and prevent data leakage. Set up Optuna or similar automated hyperparameter optimization. Improve/test Poisson and Dixon-Coles modeling parameters. Test and optimize model calibration and market-odds blending. Improve EV calculations, betting thresholds, risk management, and stake-sizing logic. Build automated champion vs. challenger testing so new models must outperform the current production model before being promoted. Create automated tests to make sure future code/model changes do not break the system. Set up Git/GitHub version control and organized development branches. Help deploy the model on a Linux mini PC for 24/7 autonomous operation. Automate scheduling, logging, error handling, backups, and model-health reports. Data Automation I also want the system to eventually collect its own data instead of requiring manual entry. This may include: Match results and fixtures xG and team statistics Sportsbook/betting-market odds Opening and closing lines Historical line movement Closing Line Value (CLV) Injury/player availability data Other useful soccer statistics This should preferably use reliable APIs when available and compliant web scraping when appropriate. The system should automatically store historical data in a structured database and update it on a schedule. Long-Term Goal The goal is to build a largely autonomous system that can: Collect Data → Update Database → Generate Predictions → Compare to Market → Calculate EV → Track Results/CLV → Backtest → Optimize → Test Challenger Models → Report Performance Future phases may include integration with permitted sportsbook/exchange APIs for automated execution, but that is not required initially. Required Skills Applicant should have solid experience with: Python pandas NumPy SciPy scikit-learn Statistical / machine-learning modeling Probability and basic statistics Time-series or walk-forward validation Preventing data leakage/overfitting Hyperparameter optimization Optuna or similar optimization libraries REST APIs / JSON Web scraping: Requests, BeautifulSoup, Selenium/Playwright or similar SQL / databases such as PostgreSQL or SQLite Git & GitHub Linux / Ubuntu Automation using cron/systemd or similar Debugging and writing automated tests Strongly Preferred Experience with any of the following is a major advantage: Sports prediction models Betting-market data Odds and implied probabilities Vig/overround removal Brier score and log loss Probability calibration Poisson models Dixon-Coles soccer models Closing Line Value (CLV) Docker Quantitative finance or algorithmic trading systems How I Want to Work I will provide the existing codebase through GitHub. Work will initially be given as small paid tasks. If the work is high quality, this can become ongoing part-time or full-time work. Please send: Your experience with Python and machine learning. A GitHub/portfolio or examples of relevant projects. Your experience with APIs/web scraping and automation. Whether you have worked with time-series validation, sports models, trading systems, or betting data. Your hourly rate. A short explanation of the difference between random train/test splitting and walk-forward validation for time-series data. Please do not apply if your experience is primarily front-end/web design with very limited Python/data-science experience.