Internal product · Knowledge systems
Coding Agent Memory Harness
Overmind indexes every coding-agent conversation your team has - searchable, shareable, and synced, with custom memory types and standing rules on top.
AI where it earns its place.
We're a product studio that designs and builds AI systems that hold up outside the demo - agents, knowledge infrastructure, specialised models, and the product design and engineering that make them usable.
We're not going to sell you AI.
AI is genuinely powerful - and it's currently being sold as the answer to a lot of problems it can't solve. We combine thirteen years of product design and engineering with hands-on AI work to find where models create real leverage in your business. When ordinary software is the better answer, we'll say so.
We mean that literally: one of our best-performing models last year was a gradient booster that trains in seconds and costs nothing to run. It's in the Lab.
What we do
AI Transformation
We find where AI actually pays off inside your organisation - and where it doesn't - then help your team adopt it incrementally, with training and retainers instead of a big-bang rollout.
Agents & Knowledge Systems
Agents are only as good as what they know and what they're allowed to touch. We build the agents, the automations, and the knowledge infrastructure underneath them - permission-aware down to the individual user.
Specialised Models
Not every problem needs a frontier model. For proven, repeated patterns, a fine-tuned self-hosted model is orders of magnitude cheaper and faster - and your data never leaves your infrastructure.
Product Design & Engineering
Design and engineering have been the studio's core since 2013. As generated interfaces flood the market, products made with attention and care stand out on sight - that's the standard we build to.
From the Lab
Experiments, internal products, and research from problems we've actually worked on. Everything labelled for what it is - no invented case studies.
Explore the Lab →
Experiment · Model engineering
Policy Classifier Agent
Prompt-defined policies - injection, exfiltration, compliance - enforced in parallel by a fine-tuned 4B, entirely self-hosted.
Experiment · Applied ML
Cinema Screening Planner
A LightGBM ranker on 1,921 screenings hit ~95% of the measurable ceiling - and proved the human scheduler was already good.
Thirteen years of evolving with software
We started in 2013 designing and building mobile products. Since then the studio has moved through full-stack development, design systems, our own products - and since 2024, the systems becoming the next layer of software.
Read our story →2013
Mobile · iOS products designed and built with care
2015
Full-stack · complete products, no dependencies
2017
Design systems · coherence at scale
2021
Our own products · our own consequences
2024
AI systems · agents, knowledge, specialised models
Prototype before promise
Every engagement starts by understanding the problem, not by picking a model. We prototype early and evaluate honestly - if the numbers say AI isn't the answer, you'll hear it in week two, not month six.
Talk to us