Side projects,
real builds.

Stuff I made on weekends and evenings to learn faster, scratch an itch, or prove a system to myself. Some have URLs. Some don't. All of them shipped, ran, and taught me something I couldn't have picked up from a course.

2 projects live in this catalogue
Built with AI-assisted dev
/ More in the workshop · check back
Catalogue

What's on the bench.

Prism SEO platform, Site Crawl dashboard
Project 01 / 2024–2025 · Multi-user web app

Prism: Full-Stack SEO Platform.

Sign in through Google Search Console, choose a property and let Prism take care of the rest: from crawling your site and auditing it for technical issues, to clustering your keywords and surfacing insights on performance, to tracking your rankings in real time and giving you a prioritised list of recommendations to work through. With AI generated insights, recommendations and alerts.

  • AI-powered keyword clustering from Google Search Console query data.
  • Smart detection of PR and brand events from irregular impression spikes.
  • Automated website crawling with technical audit reporting.
  • Content structure and health analysis.
  • Live SERP analysis identifying keyword rankings in real time.
  • AI-generated executive summaries of performance and audit findings.
  • Built-in Action Centre surfacing prioritised recommendations.
  • Integrated task kanban for workflow management.
  • Multi-user workspace functionality.
Next.jsPostgresPythonOpenAI APIGSC APIPlaywright
Discontinued · 2025 / Retired at employer's request (overlap with commercial tooling)
Financial Trading Research Platform, news terminal
Project 02 / 2024–2025 · Research platform

Financial Trading Research Platform.

A hobby-grade financial trading research platform I designed and built end-to-end in my own time using AI-assisted development. Built to learn ML, sentiment analysis, and market-data plumbing the hard way. By shipping it.

  • Real-time monitoring of social signals and news articles with NLP-based sentiment analysis.
  • Automated ingestion and scoring of financial news across multiple data sources.
  • Market data API integration for cross-referencing sentiment against price action.
  • Live dashboard surfacing sentiment trends, signal strength, and asset-level breakdowns.
  • Ensemble ML models combining gradient boosting, LSTM, and transformer architectures.
  • Technical indicator computation including RSI, MACD, Bollinger Bands, and ADX.
  • Multi-strategy execution layer covering trend following, mean reversion, and statistical arbitrage.
PythonPyTorchXGBoostFastAPIPostgresRedisHuggingFace
Research only · paused / ML training at scale gated on personal compute

/ More on the bench. Check back, or ask about anything you'd like a walkthrough of.

The point

Why this page
exists.

SEO is what I get paid for. It is not the whole picture. These projects exist because I'd rather lay the bricks myself than hand someone a deck and hope for the best. The goal isn't to prove I can code. It's to prove I can problem-solve, figure things out, and execute on my own ideas.

That instinct is the same one I bring to client work: identify the real constraint, find the lever, and build toward something that actually ships.

If anything here intersects with a problem you're working on, say hello.

Curious about any of these?

Happy to walk through the architecture, the lessons, or the parts that broke.

Start a chat →