probODds Platform

An AI-assisted personal sports-data and model project using gradient-boosted ensembles, APIs, and an operations dashboard. The architecture notes are historical; they do not establish current deployment state, predictive accuracy, or profitability.

AI-assisted personal project. Tools and implementation notes are not claims of independent professional specialization or verified live results.

Stack

Engineering highlights

• Experiments with XGBoost, LightGBM, CatBoost, Optuna, and model evaluation.
• Docker Compose services for APIs, collection, PostgreSQL, and connection pooling.
• Deployment and recovery tooling explored alongside the data pipeline; schedules and rollback guarantees require separate operational verification.
• Reverse-proxy, health-check, resource-limit, and backup configuration recorded as implementation work, not proof of reliable accounting.

Shape

Scrapers → Reader (LLM) → Quant (ensemble) → Risk → API → Next.js dashboard, over Postgres + PgBouncer