By day I'm a Principal Product Manager leading finance systems — tax technology, reconciliation, and compliance. Outside of work, I build software at the intersection of finance, automation, investing, and cloud. This site showcases those personal builds and the product thinking behind them.
A sample of product and technical work — each pairs a real user problem with a working build and the skills it demonstrates.
Problem: Tracking a portfolio and the market signals that affect it is manual and reactive — meaningful moves are easy to miss.
What it does: Screens the Russell 3000 across three risk profiles, generates per-symbol BUY/HOLD/SELL calls, synthesizes a CIO-style daily plan with an LLM, and mirrors decisions through Alpaca paper trading.
View architecture & build →Problem: Staying current across markets, news, and personal signals means scattered, time-consuming reading every morning.
What it does: A Python orchestrator gathers daily inputs from seven sources, summarizes them with an LLM, and delivers a single concise HTML digest by email on a schedule.
View architecture & build →Problem: A portfolio needs to be fast, owned, and cheap to run — without the overhead of a heavyweight CMS.
What it does: A static, CDN-served site on a custom domain with a serverless publishing pipeline for long-form and journal content.
Problem: Turning raw trip photos, voice notes, and text into a coherent, published journal is tedious and manual.
What it does: An end-to-end pipeline that ingests photos, audio, and text; auto-tags images; transcribes voice notes; and generates published blog posts on a schedule.
View architecture & build →I’m Gareth Merrifield, a Principal Product Manager focused on finance systems — tax technology, reconciliation, compliance, and the automation that ties them together. That’s my day job: defining strategy and shipping products that bring accuracy and automation to finance at scale.
Away from work, I build software for myself — Python and AWS projects across finance, automation, investing, and cloud. The Selected Builds above are those personal experiments, where I apply the same product thinking and domain knowledge to my own ideas, with APIs and AI agents.
Nineteen years across accounting, finance systems and product at Amazon, plus the applied-AI builds above. Readable in the browser or printable to PDF.
A quick ticker tape with a few indices and names I follow.