Ship it or stop
If a model cannot reach production within a defined timeline, we say so early and refund unspent fees. Research experiments belong in universities. Our job is to deliver working software.
How we got here
Logica AI Masters was founded in 2021 by two machine-learning engineers who had spent a combined twelve years at large consultancies watching proof-of-concept models gather dust. The pattern was always the same: a flashy pilot, enthusiastic stakeholders, then silence once the contract ended because nobody had planned for deployment, monitoring or retraining.
We registered the company in Schmittridge, England, and took on our first client within a month. That project was a demand-forecasting model for a mid-sized food distributor with 14 warehouses. The model went live in six weeks and is still running today, retrained monthly on fresh order data.
Since then we have completed 14 projects across logistics, healthcare, retail and legal services. Every single model we have built is either still in production or was deliberately retired when the client changed business direction. None has been abandoned.
What we believe
Four principles that shape every engagement.
Explain the limits
Every model has failure modes. We document them in plain English and share them with your team before launch. A model that surprises its operators is worse than no model at all.
Own the data conversation
Bad data produces bad predictions. We audit data quality at the start of every project and will push back if the training set is too small, too biased or too stale to support the use case.
Fixed scope, fixed price
Open-ended retainers reward slow delivery. We agree on a scope, a price and a timeline before writing a single line of code. If we underestimate the effort, that is our problem, not yours.
The team
Five people, each with a specific role. No account managers, no overhead.
Daniel Carr
Co-founder, ML engineering
Eight years building production ML pipelines. Previously at a London fintech where he deployed fraud-detection models processing 2 million transactions daily.
Priya Anand
Co-founder, data science
PhD in computational linguistics from Edinburgh. She designed the NLP pipeline that a regional NHS trust now uses to triage patient correspondence.
Sam Okafor
ML engineer
Joined in 2023 after three years at a computer-vision startup. He handles image-classification and object-detection projects from labelling strategy through to API deployment.
Leah Thornton
DevOps and infrastructure
Leah containerises our models, sets up CI/CD pipelines and configures monitoring dashboards. Before joining she managed Kubernetes clusters for a SaaS company in Manchester.
Milestones
Key moments since we started.
March 2021
Company registered in Schmittridge, England. First client signed within four weeks: a food distributor needing demand forecasting across 14 warehouses.
September 2021
Delivered our first NLP project: an email classifier for a legal firm that reduced manual sorting from three hours per day to twenty minutes.
June 2022
Completed a computer-vision quality-control system for a packaging manufacturer. The system inspects 800 items per hour on a single camera.
January 2023
Sam Okafor joined as our third engineer, bringing deep experience in image classification. We expanded into healthcare imaging projects the same quarter.
October 2023
Launched the ongoing-partnership plan after clients asked for continuous model retraining. Three clients signed in the first month.
April 2024
Leah Thornton joined to handle infrastructure full-time. Average deployment time dropped from five days to one and a half.
2025
Fourteen projects completed, zero abandoned models. We are now accepting two new engagements per quarter to maintain delivery quality.
Want to discuss your project? Call us at +44 55 4350 4643 or email [email protected]. Our office is at 17 Christiansen Glade, Schmittridge, England, QQ83 1TC, United Kingdom.