About us
Why we started this
We started Shrink Ray because we see the same thing in different places: capable operations, run by people who know what's wrong, working around problems that are entirely solvable.
We want to build a future where the tedious parts of running an organization take care of themselves, and people get to do the part that truly requires a person.
Who you'll work with
Andrew Larimer
Co-founder
Andrew leads Gen AI and ML engineering curriculum for Google's Partner Learning Services, where he develops and approves the training Google's partner engineers use to learn its newest AI tooling.
Before that he spent four years at SpringML as a machine learning practice manager, where he designed the system a US state used to track provider inventories and distribute COVID-19 vaccine from the CDC through local health districts, built a hardware-to-cloud pipeline moving image data off a fleet of municipal vehicles for pothole and blight detection, architected a BigQuery warehouse consolidating a state HR department's data sources, and led research and implementation of the firm's first generative AI language work. He also wrote their Guide to Effective Engineering.
He holds a master's from UC Berkeley's School of Information and is a Google Authorized Trainer. Before engineering, he built and sold a New Orleans media company. His undergraduate degree is a BFA in drama from NYU, which turns out to be more useful for explaining technical systems to non-technical stakeholders than it sounds.
Ryan Murphy
Co-founder
Ryan builds the data systems and analytics that leadership uses to make decisions. Ryan came to Shrink Ray from Stanford Health Care, where he worked on Workday finance and supply chain integrations: certified in Workday integrations, implementing AI tooling inside the platform, reporting, and automating stretches of the procure-to-pay process that had been running on people.
Before that, two years at LCMC Health, an eight-hospital system, as decision support and supply chain analyst. He implemented the Prism layer connecting Workday to outside data sources so operations could act on numbers while they still mattered, and the Python reporting behind supply chain KPIs and surgical inventory. Before that, three years as senior data analyst at Achievement First, a 41-school charter network, where he built the reporting leadership ran on out of AWS Redshift and the models on top of it: multivariate and logistic regression, decision trees, Loess regression, predicting assessment outcomes and identifying what drove student performance. For Tulane's surgical residency program he built the system that scores personal statements and letters of recommendation against the program's own rubric, which predicted applicants' post-interview outcomes more closely than the screen it replaced.
He has a law degree from Boston University with a concentration in financial law and a philosophy degree from the University of Georgia, taught computer science, and spent three years training teachers across Louisiana to deliver computer science curriculum. He's spent his career explaining difficult concepts to people who need answers and background, which is a big part of this job.
Cath Larimer
Designer
Cath leads product design at Calm. Before that she spent four years at The New York Times, most recently as a lead product designer, working on the interfaces millions of people use without ever thinking about them, which is the point.
Her freelance work runs through blockchain, B2B services, and media, including information architecture and journey mapping for an enterprise AI platform. Earlier she spent three and a half years at Thackway McCord as lead UI/UX designer, doing product design alongside branding and corporate identity work for clients in the US and UK, and did stints at BCG Digital Ventures, Publicis Sapient, and Huge. She started as a graphic designer for L'Oréal and philosophy, and before that as a film editor, cutting more than 300 videos for clients in fintech, media, and nonprofits.
The reason a design person is on an engineering team: most automation fails at the handoff. A system can be technically correct and still produce something a human has to fight with. And if the review step takes as long as the work it replaced, nothing was saved. Cath makes sure the thing that reaches a person is worth the time it takes them to read it.
Where we are
We're based in New Orleans. Most of the work happens remotely; when it matters, we show up.
What we don't do
We're not the right firm for everyone.
- We won't take on work at a price where we don't believe we can do it properly. But we're glad to build a pilot within your budget so you can see the value, and then talk about expanding from there.
- We're not going to try to sell you a more technically advanced and expensive solution than what you need. And we'll tell you when what you're describing doesn't need AI at all.
Let's talk
Let us know if you've got something that needs shrinking.