Our verdict
Fuzzy Labs is an AI development, implementation and consulting firm in Manchester, England, that Tom Stockton and Matt Squire lead. LinkedIn lists it as founded in 2019 with about 25 employees, and its About page names more than 20 people on the team, including engineers with PhDs in astrophysics, particle physics and machine learning (Fuzzy Labs; LinkedIn).
The firm sells open-source MLOps work: building the pipelines, experiment tracking, data versioning and deployment systems that take a model from a notebook into production, then hand the stack to the client’s own engineers. Its published case studies include an MLOps pipeline for behavioural AI company zally that cut model deployment from two days to five minutes, and a production support assistant for a major US sports broadcaster that cut average first-response time from 64 minutes to 37 seconds (Fuzzy Labs case study).
For a team that already has data scientists and wants production infrastructure without a proprietary MLOps vendor lock-in, Fuzzy Labs is a small specialist that builds with tools the client will own.
The right buyer
- Companies with an existing data science team that want open-source MLOps infrastructure built and handed over.
- Teams that need model deployment, experiment tracking and data versioning without a proprietary platform lock-in.
- Buyers who want engineers embedded beside their own staff, as in the zally engagement.
Reasons to look elsewhere
- We found no independent review platform coverage; available client quotes appear in Fuzzy Labs case studies.
- It does not publish pricing or engagement sizes, so buyers who need a published rate card should look elsewhere.
- With about 25 people it is a specialist MLOps shop, not a large delivery bench for broad GenAI product builds.
Scores
| Criterion | Weight | Score | Basis |
|---|---|---|---|
| Production track record | 20% | 7.2 | Named zally results and a published sports-broadcaster support assistant; other clients shown as logos only. |
| Technical depth | 15% | 7.8 | Open-source MLOps depth, including its Matcha starter kit and ZenML-centred pipelines. |
| Industry depth | 10% | 6.4 | Work spans behavioural AI, broadcast media, energy and public-sector policing; no single vertical practice. |
| Delivery model and handoff | 15% | 7.6 | Embeds as an extension of the client team and designs for independent ownership after handover. |
| Commercial clarity | 10% | 5.0 | No pricing or engagement minimum published. |
| Platform independence | 10% | 8.6 | Open source by design; clients own the solution rather than a vendor platform. |
| Stability and scale | 10% | 6.6 | Founded 2019; about 25 staff. |
| Independent evidence | 10% | 6.2 | Evidence is company case studies and named client quotes; we found no Clutch or similar review page. |
| Overall | 100% | 7.0 | Weighted average |
Where Fuzzy Labs fits
| Need | Fit | Why |
|---|---|---|
| MLOps pipeline and model deployment | Strong | Core offer; zally case. |
| Productionising an internal AI assistant | Strong | US sports broadcaster case. |
| Open-source LLMOps for developers | Good | Published LLMOps case work. |
| Large multi-year systems-integrator programs | Weak | Small specialist bench. |
What clients and leaders say
“We’re building a new type of AI here at zally and I need the best minds to work with the best tools in order to achieve our goal. Fuzzy Labs provided those tools, no question.”
“Fuzzy Labs built the ‘racetrack’ to enable our data science and ML engineers to build and deploy models at scale.”
“What really stood out was how generously they shared their knowledge. They took the time to explain concepts clearly, answered all my questions with patience, and really helped me upskill along the way.”