Our story
Weird Fishes Network brings together a team of forty across London and Manchester. We build practical machine-learning systems that help organisations make better decisions.
Turning data into decisions.
Weird Fishes Network builds machine learning systems for retailers, banks, insurers, hospitals, utilities and more. We bring data, technology and domain expertise together to help teams make confident decisions.
Models running in production, by quarterAcross selected engagements
0
team members
0
projects delivered
0
sectors served
0
models in production
0
data processed
Where the work went
Retail and banking were our largest practices, but some of our most interesting problems came from water networks, hospital wards and telematics boxes. Every engagement started with one question: what decision will this model change?
Work across eight sectors and a wide range of real-world decisions.
Our work
Explore selected engagements to see the challenge, the data, the modelling approach and what changed for each client.
Case studies are anonymised and representative of engagement types; client identities are withheld.
Products
Alongside consulting, we build software that turns repeated, complex work into practical tools for teams.
How we worked
Our projects follow a clear path from raw data to a monitored model in production. The pipeline below brings together the practices we use across engagements.
Engagement model
1
Discovery
Two weeks with the people who would use the model. We mapped the decision, the data and the cost of being wrong.
2
Proof of concept
Six to eight weeks on historical data, with a backtest the client's own analysts could rerun.
3
Production
Pipelines, APIs and dashboards built into the client's stack, with alerts on drift and data quality.
4
Handover
Documentation, training and paired support until the internal team owned the model end to end.
Technology we used
Contact
For business enquiries, partnerships or questions about our work, please get in touch.
General enquiries
Replies may take a few days. We cannot share client data or materials covered by confidentiality agreements.
Our team
A multidisciplinary team working across data, engineering, product and client partnerships:
Founder and MDHead of Data ScienceHead of EngineeringPrincipal Data Scientists (4)Senior Data Scientists (9)ML Engineers (7)Data Engineers (8)Product Managers (3)Designers (2)Client Partners (4)