Productised Services

Fixed-scope engagements for AI platform engineering. Clear deliverables, no long sales cycles.

Architecture Review

90 minutes

Bring your use case, and I'll map the path from ingestion through OCR, retrieval, extraction, eval, and rollout.

Deliverables

  • System architecture diagram
  • Risk assessment and mitigation strategies
  • Cost model and scaling projections
  • Implementation roadmap with next steps
£450
one-time
Book Review

Two-Week Pilot: Private Doc-Intelligence on K8s

10 business days

Fixed-price pilot. We deploy a private document intelligence pipeline in your environment with clear SLOs.

Targets

  • ≥ X docs/min throughput
  • Precision/recall ≥ Y/Z on gold set
  • p95 latency ≤ N seconds
  • Cost ≤ £C per 1k docs

Deliverables

  • Helm charts for deployment
  • Grafana dashboards
  • Rollback and disaster recovery plan
  • Security assessment and notes
Fixed fee
contact for quote
Request Quote

Ops & Evaluation Retainer

ongoing

Monthly retainer covering SLO monitoring, drift detection, gold-set curation, red-teaming, cost optimization, and incident response drills.

What's included

  • SLO monitoring and alerting setup
  • Model drift checks and retraining triggers
  • Gold-set curation and evaluation runs
  • Red-team testing for hallucinations and safety
  • Cost analysis and GPU utilisation optimisation
  • Quarterly incident response drills
Monthly
contact for pricing
Discuss Retainer

Frequently Asked Questions

Do you work with existing infrastructure?

Yes. I work within your existing Kubernetes clusters, CI/CD pipelines, and security policies. If you're on AWS, GCP, Azure, or on-prem, we'll adapt to your environment.

What if the pilot doesn't hit targets?

Targets are agreed before starting. If we don't hit them, you don't pay the fixed fee. Simple as that.

Can you help with compliance (GDPR, HIPAA, etc.)?

Yes. Private deployments with no data leaving your VPC, audit logs, encryption at rest/in-transit, and SBOM generation are part of the standard approach.

Do you provide ongoing support after the pilot?

That's what the Ops & Evaluation Retainer is for. Or, you can take the delivered Helm charts and run it yourself—full ownership, no vendor lock-in.

What technologies do you typically use?

Kubernetes, Helm, Docker, NVIDIA GPUs with MIG, Prometheus, Grafana, LangChain/LlamaIndex (when appropriate), FastAPI, Python, Go. The right tool for your constraints, not the latest hype.

Ready to start?

Book a 30-minute Architecture Review to map out your AI platform needs.

Book Now