Skip to main content

Paul Sellars · Principal AI and Cloud Platform Architect

Inventor on UK Patent Application GB2519240.2

What I do

Building the platforms, models, and integrations that move regulated enterprises through the AI transition.

Sixteen years across cloud, data, and applied machine learning. First-class BSc in Artificial Intelligence (2010) — fifteen years of applied AI experience before the recent commercial inflection.

How I work

Force multiplier, not feature factory.

I ship, selectively. Platform tooling, IaC modules, the integration work that ten downstream features depend on, the architecture decisions that quietly shape every later choice: that’s where architect attention pays back. The best engineering investment I can make for a client is rarely a single feature; it’s the system that lets their team ship confidently for years afterwards.

Anything the team can carry stays with the team, by design. Permanent engineering teams own the feature backlog; the architect bottlenecking on every ticket is the failure mode I’m there to prevent. Engagements are built for handover continuity — no single-person dependencies, work that survives my departure, and coaching toward independent problem-solving so the engineering organisation gets stronger, not more dependent on a single architect.

Focus is regulated industries — energy, utilities, financial services, healthcare, insurance, defence, pharma, public sector — where AI and data systems must integrate with legacy infrastructure, meet safety and compliance standards, and operate at scale. What carries across is the operating experience in environments where audit, legal, and risk get a vote, not the specifics of any single vertical. New compliance regimes (HIPAA, FINRA, SR 11-7, FDA, Solvency II) get learned as part of the engagement.

  • Pick the right work

    Architect attention goes to the work ten downstream features depend on. Tickets the team can carry stay with the team.

  • Continuity by design

    No single-person dependencies. Product maintainability outlives the engagement.

  • Senior decisions in the open

    Architectural judgment on the hard calls, with the reasoning visible so the team can carry it forward after the engagement ends.

Focus areas

Where AI, data, and platform engineering meet operational reality.

Strategic AI & Data

AI architecture for regulated industries. Data platform engineering, ML systems design, and AI commercialisation in production.

Cloud Platform Engineering

AWS-native architecture, Kubernetes, infrastructure-as-code, serverless-first design, and secure CI/CD. Cost and reliability built in, not bolted on.

Applied ML & AI Systems

Production machine learning, model and inference architecture, and the integration patterns that put AI to work inside existing operational stacks.

Selected work

Engagements at scale, across regulated and high-traffic environments.

Current engagement

Principal AI and Cloud Platform Architect, PS Digital Limited

September 2023 to present

Independent strategic architecture engagement for clients in regulated industries. Architectural authority and de facto programme leadership across cloud, data, and AI workstreams; permanent engineering teams own feature delivery for continuity beyond the engagement. Designs the platform tooling, governance patterns, and infrastructure-as-code modules that enable distributed teams to deliver at scale.

  • Architected the data platform and supporting systems for the client engagement
  • Authored job specifications and selected the DevOps lead and contract delivery team
  • De facto programme manager across multiple engineering workstreams
  • Set architectural direction and identified gaps; permanent team owned feature implementation
  • Architectural review and decision authority across cloud, data, and AI
  • Coaching engineering team toward independent problem-solving and self-directed research
  • Designed AWS Control Tower governance and multi-account architecture
  • Built infrastructure-as-code modules and migration patterns for self-serve cloud modernisation
  • Established modern CI/CD using Terraform, AWS CodeDeploy, AWS CodeBuild
  • Delivered applied machine learning and AI systems in production for regulated industry use cases
  • Inventor on UK Patent Application GB2519240.2
UK Patent, public record

UK Patent Application GB2519240.2

Application
GB2519240.2
Filed
17 November 2025
Status
Filed
Inventor
Paul Sellars (on the application)
Engagement

Start a conversation.

Inquiries welcome from organisations across regulated industries — energy, utilities, financial services, healthcare, insurance, defence, pharma, public sector — where AI and data architecture meet compliance and operational safety constraints. Comfortable onboarding to unfamiliar compliance regimes (HIPAA, FINRA, SR 11-7, FDA, Solvency II, and adjacent) as part of the engagement. Open to senior architecture engagements that originate the vision AND senior IC roles that land a defined architectural vision in production at scale: principal architect, fractional CTO / CDO, cloud architect, Head of Platform Engineering, transformation programmes, and senior IC seats on complex production rollouts.

Or email ping@psdigital.dev