What I do for companies
I architect and build backend and AI systems: integrations, event-driven services, data pipelines, and the models that run on them. Most of my work is with teams who need a senior engineer to own a hard part of the product rather than staff a seat.
How we can work together
End to end product build
Zero to one. Architecture, prototyping, implementation, QA, release, and the maintenance after it. I own the outcome rather than a task on it.
Embedded engineer
I join an existing team and codebase as a senior engineer for the length of the work, in your tooling and at your cadence.
Architecture and audits
Architecture review, performance and code audits, and a senior second opinion before a decision becomes expensive to reverse. Short engagement, clear deliverable.
AI and machine learning
Forecasting, anomaly detection, and LLM features built into production systems rather than bolted onto them, with the data pipelines to keep them fed.
Unity and game technology
Engine-level work, editor tooling, and simulation systems for teams building in Unity. The engine I ship my own games in.
Selected work
Client names withheldMy client work is covered by confidentiality, so these are kept general: the sector, the problem, and the shape of the result. Happy to go deeper in a conversation, within what I am able to share.
Subscription SaaS
A multi-tenant platform keeping operational and financial data in sync across third-party business systems.
Lead backend developer and system architect
- Native integrations with third-party business systems over REST and webhooks, released through CI/CD without downtime.
- A modular configuration engine that moved the large majority of client setups to self-serve, cutting onboarding from days to hours.
- A bidirectional sync engine that removed manual data entry and held connected systems consistent in real time.
- Anomaly detection and predictive analytics over live data streams, surfacing discrepancies before they reached customers.
- A rebuild of the architecture into modular event-driven services, improving scalability and fault tolerance.
Rebuilt a core reporting pipeline and cut its runtime by roughly an order of magnitude through query optimisation and stream processing.
- TypeScript
- Node.js
- Docker
- Kafka
- MongoDB
- PostgreSQL
Enterprise retail
Data and cloud infrastructure for a high-volume retail operation.
Software engineer and lead AI solution architect
- Migration of a legacy monolith to microservices using domain-driven design, consolidating real-time data from a large distributed estate.
- Machine learning models for inventory management and demand forecasting, including time-series models, cross-validation pipelines, and automated retraining.
- High availability through database sharding and replication, with OAuth-based authentication across services.
- Coordination across machine learning engineers, developers, and analysts to keep live data processing fault tolerant at scale.
Demand forecasting measurably reduced stock written off and contributed a revenue lift across the network.
- Python
- scikit-learn
- pandas
- MLflow
- Kubernetes
- Kafka
- PostgreSQL
Simulation and training
An immersive training platform built to generate unlimited live scenarios.
Lead simulation developer
- End-to-end backend for the simulation platform, built in Unity with C#.
- Scalable architecture covering data persistence, logging, and modular scenario generation.
- Simulated compromised environments that raised the realism and operational relevance of exercises.
- Requirements gathered and documented directly with the client at every sprint.
The modular scenario architecture produced a substantial measured improvement in trainee performance.
- Unity
- C#
Clinical imaging
An anomaly detection tool for diagnostic imaging, and training for the people maintaining it.
AI consultant
- Guided development of an anomaly detection model for diagnostic images to a high precision threshold.
- Reduced diagnostic error and shortened report turnaround in time-critical settings.
- Delivered a practical training programme on applied AI for engineers and clinicians.
- Python
- Computer vision
Stack
Backend
- TypeScript
- Node.js
- Microservices
- Event-driven architecture
- REST APIs
- Kafka
- Docker
- Kubernetes
- AWS
- CI/CD
Data
- PostgreSQL
- MongoDB
- OpenSearch
- Neo4j
- Sharding and replication
AI and machine learning
- LLM integration
- Forecasting
- Anomaly detection
- NLP
- scikit-learn
- pandas
- MLflow
Games and simulation
- Unity
- C#
- Editor tooling
- Physics systems
- Shader programming
Mobile
- iOS
- Swift
- SwiftUI
- In-app purchases
- App Store release
Practice
- Software architecture
- Domain-driven design
- Agile
- Code review
- Mentoring
Tell me what you are building
Whether it is a system to design, a team to join, or an architecture you want a second opinion on. You would work directly with the person writing the code.
Start a conversation