What we can build for you
Depth across
the whole system.
A model is one part of the picture. We can help with the software, data and infrastructure that turn it into a useful capability.
AI agents & LLM systems
Agents and LLM workflows that use your own data and tools: retrieval-augmented generation (RAG), natural-language access to databases, tool access through the Model Context Protocol (MCP), fine-tuning of open models and private deployment. We build them with the same engineering discipline as the rest of your software.
- AI agents & MCP
- RAG & fine-tuning
- Private LLM deployment
See: LLM-assisted reports in Platform modernisation
Applied ML & algorithms
Models and algorithms grounded in mathematics: prediction on small and imbalanced data, multi-armed bandits and simulation for bidding, graph matching and recommendations. We choose the method that fits the problem.
- Predictive models
- Optimisation & bandits
- Graph matching & simulation
See: Predictive learning analytics, Real-time bidding optimisation, Planogram compliance, Enterprise data warehouse
Computer vision & AR
Detection, optical character recognition (OCR) and feature tracking in shelf photos, on packaging and in live video, supported by tools for manual and automatic annotation and for training. A coffee machine augmented reality (AR) demonstration in the computer vision platform case processed a 40-60 fps video stream.
- Detection, OCR & tracking
- Real-time video & AR
- Manual & auto-annotation
See: Computer vision platform, Planogram compliance
Data engineering
Pipelines and warehouses that make data usable: ETL and ELT at scale, analytical databases, change data capture (CDC) and multi-region data that respects residency rules, designed with cost and latency in mind.
- BigQuery, ClickHouse & PostgreSQL
- Airflow & Kafka
- CDC & data residency
See: Enterprise data warehouse, Real-time bidding optimisation, Platform modernisation
Platform modernisation
From a monolith tied to one cloud to services designed to run with different cloud providers or on your own servers: architecture, infrastructure as code, GitOps delivery and lower running costs.
- Go & Python services
- Terraform & GitOps
- Cloud migration & Kubernetes
See: Platform modernisation