— About

We build the agents we'd want to work with.

Hyyve is a small studio. We started in 2024, after spending years on the inside — building ML platforms, scaling operations teams, trying to ship AI inside companies that weren't built for it. We saw what worked, and what didn't, and wanted to do it differently.

— How we work

The work, then the tech.

We don't lead with models or frameworks. We lead with the actual job to be done — and we pick tools that fit, not the other way around.

Two-week loops, no demos.

Real users, real data, real tools from week one. If something can't survive contact with the team, it isn't real.

Small teams, high context.

Three people who know your operations cold will out-ship a team of fifteen who don't. We staff to context, not to invoice.

Honest about what AI can't do.

We've turned down work where the right answer was a spreadsheet. We'd rather lose the project than oversell the technology.

— The team

Sasha Reyes

Co-founder · Strategy

Former head of AI ops at a Series D logistics company. Spent two years finding out what doesn't work.

Devraj Mehta

Co-founder · Engineering

Ten years in ML platform infra. Believes most AI failures are integration failures wearing a costume.

Linnea Brask

Principal · Design

Designs the operator-facing surfaces of every agent. Thinks about exception queues more than is healthy.

Wole Adeyemi

Engineering Lead

Builds the tool servers and the eval suites. The reason our agents don't quietly drift.