Hamilton Ulmer

👋 Hi – I'm Hamilton.

I’m a data scientist and product developer. I combine statistical & machine learning skills with product thinking, engineering, and UX & UI design to make magical, useful things.

I’m currently the head of Applications Engineering at Rill Data. We’re currently focused on building Rill Developer, a new tool that makes it effortless to transform your datasets with SQL.

Prior to that, I was at Mozilla. Some selected projects from my time there:

The Iodide Projectbringing data science to the web through Web Assembly – Iodide was an experiment with making a radically reader-centric web-based notebook environment. I co-founded the project and developed it with a few colleagues. Our ambition was to compile all the major scientific computing stakes to WebAssembly and replace the computational kernel with a CDN. We shuttered the project after Mozilla’s first round of layoffs, but I’m still bullish on the possibilities. The Pyodide project, which was the Python WebAssembly stack that included NumPy / Pandas / Matplotlib / Scipy, now lives on in Jupyterlite.

Mozilla Rallya first-party data market that enables people to transact their browsing data equitably and ethically – A recent Pew poll found that more than 31% of Americans are “almost constantly online”, and that the overwhelming majority are on several times a day. That should probably not be surprising to you – if you’re reading this, you’re probably “extremely online” yourself. For better or for worse, browsers are a window into our lives. We work on them, we play on them, we’re distracted by them. And because of that, there’s incredible economic value in data about our browsing. Rally gives people a way to directly profit from the data they generate. I helped lead product development & engineering, dataset design, customer discovery, prototyping, and strategy, something of a one-man startup utility player.

Metrics-Graphics and Graph Paperholistic data visualization & design systems – one thing I’ve learned throughout my career is that polish helps build trust in interfaces, especially with data-centered decision support interfaces. On its own, data visualization is not enough. Metrics-Graphics (7.4k stars on Github) was an early attempt at producing an opinionated design system for business charts, utilized at Mozilla and many other companies (I handed off maintenance a few years ago). Graph-Paper is a similar concept, built entirely in Svelte, which enables data workers to build elegant, reactive data interfaces. It currently powers a number of data tools at Mozilla today.

Data Tools and Decision Support Systems – with others, I designed and built a number of foundational data tools and interfaces to help executives and product / engineering leaders make better decisions with data. Data-centric UX is a double-challenge; you have to solve problems of trust (can we trust data as a concept? can we trust these data?) and comprehension (what is this interface telling me? how do I make a decision with these data?). Building tools and workflows is one way to scale up data expertise.

Prior to this career, I was a musician and arranger, working with artists big and small in Nashville and Los Angeles. Ask me in person about it. I’ve retired from the industry (for now) – a rollercoaster of ups and downs that I wouldn’t trade for anything.