A spreadsheet error in one of the most influential economics papers of the decade helped justify austerity across Europe and the US. Millions lived with the consequences. The error was only found years later, by sheer luck, long after the damage was done.
Flawed analysis shapes policy, investment and lending every day, costs billions every year, and mostly goes unnoticed. AI makes it worse: rigorous-looking work is now cheap to produce.
Refine exists to catch these errors before anyone acts on them. We're building the infrastructure that verifies high-stakes work; think CI/CD for everything beyond code. It reviews a document like a team of experts, checking the maths, data, logic and argument, in minutes.
The best AI review system. Refine 5 beats frontier models such as GPT-6 Astra in over 90% of blind head-to-heads across ten fields, from economics to biology. It emulates expert human judgment, coordinating hundreds of calls to frontier models so none of their blind spots get through. We tune it on documents seeded with the errors frontier models miss.
The de facto AI reviewer for economics. Econometrica and the AER run every paper they publish through Refine. It's used at 24 central banks and by researchers at Oxford, MIT, Stanford, Yale and Cambridge; Cambridge, UCLA and NYU provide it to theirs.
Beyond academia. 29,000+ reviews sold at 1,000+ institutions, including litigation consultancies, hedge funds and companies bidding on public tenders.
Enterprise-grade. ISO 27001 certified, SOC 2 Type 1 (Type 2 in progress).
Where we are. Seed stage: fewer than 10 people, $5–10M raised, and revenue in the $1–10M range.
We build AI systems that beat frontier models at specific tasks. Refine 5 is one of them.
You'll help us design and improve these systems. The job is getting as much as possible out of today's frontier models.
What mattersA very good mental model of what frontier models can and can't do. The kind you only get from spending a lot of time with them. You should be obsessed with AI, and with what it consistently gets wrong.
Excellent intuition for using LLMs, above all inside a system. Not just one clever prompt; how many calls fit together to get the most out of them.
You should know the basics (MCP, APIs, tool use), but those are learned fast.
Who you areNot necessarily a conventional ML engineer. Credentials don't matter much, and you don't need a theoretical background.
You've spent a lot of time experimenting with LLMs and got unusually good at it. Maybe you've built pipelines, harnesses or domain-specific tools. Maybe you've found ways of working with AI that let you do things you couldn't do yourself.
You can code well enough to prototype and test your own ideas.
You're experimental and inventive. Talent and output beat a conventional background.
How to applySend your CV or LinkedIn, plus something you've built with LLMs and a few lines on why it beats the obvious approach.
Similar Jobs
What you need to know about the Colorado Tech Scene
Key Facts About Colorado Tech
- Number of Tech Workers: 260,000; 8.5% of overall workforce (2024 CompTIA survey)
- Major Tech Employers: Lockheed Martin, Century Link, Comcast, BAE Systems, Level 3
- Key Industries: Software, artificial intelligence, aerospace, e-commerce, fintech, healthtech
- Funding Landscape: $4.9 billion in VC funding in 2024 (Pitchbook)
- Notable Investors: Access Venture Partners, Ridgeline Ventures, Techstars, Blackhorn Ventures
- Research Centers and Universities: Colorado School of Mines, University of Colorado Boulder, University of Denver, Colorado State University, Mesa Laboratory, Space Science Institute, National Center for Atmospheric Research, National Renewable Energy Laboratory, Gottlieb Institute



