Everything we will actually take on.

LLM pipeline engineering

Multi-step orchestration, structured workflows, YAML-driven prompt configuration, AWS Bedrock and equivalent platforms.

Semantic search & RAG

Retrieval, reranking, HyDE, query expansion and rewriting, domain taxonomy and vocabulary layers, GPU-served at low latency.

AI reliability & evaluation

Benchmark datasets, test harnesses, automated regression gates, replay proofs and written, scored accuracy audits.

Model fine-tuning

Domain classifiers, taggers and custom embedding models when a general-purpose model is not accurate enough on your data.

AI features in existing products

Search, extraction, summarisation, classification and chat, retrofitted into software that already has real users.

Explainable & auditable AI

Output provenance, grounding verification, prompt versioning and quality gates for regulated and high-scrutiny environments.

Web applications

SaaS platforms, admin dashboards, internal tooling, customer portals, marketplaces, booking systems and B2B workflow apps.

• Frontend: React, Angular, Vue, Next.js, TypeScript, PrimeNG, Tailwind
• Backend: Python/FastAPI, Node/NestJS, Ruby on Rails, Laravel, Java/Spring
• API design and contract collaboration

Mobile applications

Native and cross-platform apps taken through build, review and release. Over 40 published apps across the team.

• Native iOS (Swift, SwiftUI) and Android (Kotlin, Java)
• React Native, Flutter, Ionic
• Offline-first PWAs, service workers, responsive design
• App Store and Play Store publishing and release management

System architecture & design

Devising, configuring and operating the right combination of systems and components to hit a real business objective — not a diagram for its own sake.

Legacy modernisation

Taking over an existing codebase nobody wants to touch, understanding it properly, and evolving it without breaking the business running on it.

AWS engineering

EC2, ECS, EKS, Lambda, S3, RDS, DynamoDB, CloudWatch, IAM, VPC, CloudFront, ECR and CodePipeline in production.

CI/CD pipelines

GitHub Actions, Codemagic and mobile release pipelines that reliably produce a signed build and ship it to TestFlight or Play.

Infrastructure as code

Terraform, CloudFormation and scripted infrastructure that can be torn down and rebuilt from a repository with confidence.

GPU & ML infrastructure

Model serving on GPU instances, container images, environment replication between staging and production.

Migrations & upgrades

Database and platform upgrades, cloud migrations and version end-of-life moves executed without downtime.

Performance optimisation

Slow endpoints, expensive queries, runaway cloud bills and systems that fall over under load — measured, then fixed.

Blocked releases

Dead CI pipelines, builds that will not sign, store rejections and dependency hell — diagnosed to root cause, not worked around.

Production firefighting

The intermittent bug nobody can reproduce, traced through real production logs and closed properly.

Security incident response

Compromise diagnosis, malware and web-shell removal, clean restore, credential rotation, hardening and a written incident report.

Authentication & access control

Auth0, SSO, OAuth 2.0, JWT and role-based access, plus finding and closing authorisation flaws in what you already have.

Regulated data

Healthcare and HIPAA-adjacent systems, fintech, and payment-adjacent flows — built by people who have worked under that scrutiny before.

Backup & disaster recovery

Backup strategy that is actually tested, so the restore works on the day you genuinely need it.

Payments & enterprise

Stripe, PayPal and VISA gateway integrations, plus enterprise systems including Oracle, IBM middleware and CRM platforms.

Event-driven & real-time

Webhooks, queues, Kafka, Redis background processing, WebSocket live sync and geospatial data with PostGIS.

Testing & QA

Test automation, benchmark suites and release verification — including the AI-specific evaluation work most QA teams cannot do.

Architecture review & due diligence

An honest technical assessment of a codebase or a system, whether it is yours or one you are considering acquiring.

Fractional CTO

Engineering leadership, hiring input and technical decision-making for companies not ready for a full-time CTO.

Technical documentation

Architecture documentation, runbooks and handover material written so the next engineer can actually use it.