01
Production Weather CLI
Turn a fragile script into a resilient command-line tool, with exponential backoff, three-layer input validation, and structured logging.
Projects
These are the systems I designed, built, tested, and deployed while creating Python Developer: From Requests to Production. Together they show the path from a resilient script to production services on AWS.
01
Turn a fragile script into a resilient command-line tool, with exponential backoff, three-layer input validation, and structured logging.
02
A personal Spotify listening archive. OAuth 2.0 login, a SQLite schema for track snapshots and raw Spotify payloads, then a Flask dashboard for Forgotten Gems and listening analytics.
03
Move the Music Time Machine from SQLite to PostgreSQL, with JSONB queries over raw track payloads, a connection pool, and updated dashboard queries.
04
Build your own REST API with FastAPI, combining several news sources behind one API, with API key authentication, rate limiting, and auto-generated OpenAPI docs.
05
Containerise the API with Docker, deploy it to AWS on ECS Fargate with an RDS database, and add a GitHub Actions pipeline that deploys every push.
06
The capstone. A production API that pulls from NewsAPI, the Guardian and Reddit, with PostgreSQL, API key auth and rate limiting, deployed on AWS with monitoring and CI/CD, plus one extension such as sentiment analysis.