Projects

These projects show how I think about AI-native software delivery: not as isolated coding assistants, but as governed systems with context, permissions, evidence, evaluation, and human approval.

Pathfinder Programme

Pathfinder is a governed AI-native SDLC lab. It explores how software teams can turn messy delivery intent into formal specs, compare agent frameworks, orchestrate human-agent collaboration, capture evidence at every control point, evaluate agent behaviour, and feed delivery outcomes into a live risk radar.

The first phase focuses on the agent framework landscape because framework choice shapes the rest of the operating model. The question is not “can this framework create an agent?” The question is “can this framework support governed software delivery?”

Primary Portfolio

• Agent Framework Landscape Lab - comparing MAS frameworks against real SDLC automation criteria.
• Multi-Agent SDLC Control Plane - the anchor reference architecture for governed agent-assisted delivery.
• Agentic Governance Engine - converting engineering standards into enforceable agent and CI quality gates.
• Context Quality Benchmark Harness - measuring whether software agents receive faithful, relevant, useful context.
• Engineering Health Radar - delivery, quality, risk, flow, and governance signals for engineering leaders.

Upcoming Build Tracks

• SterlingPay Company-in-a-Box - a synthetic fintech SDLC substrate with realistic strategy, backlog, PR, CI, incident, and delivery data.
• Intent-to-Spec Engine - turning messy requirements into structured specs, acceptance criteria, constraints, risks, and traceability.
• Agent Delivery Workbench - coordinating specialist agents through spec, architecture, QA, security, implementation, review, and evidence stages.
• Governance Evidence Graph - moving beyond audit logs into evidence-generating governance and outcome learning.
• EvalOps Harness - testing agent workflows for spec conformance, policy compliance, red-team resistance, and regression risk.

Match This Work To A Role

Hiring for an AI Engineering Manager, platform engineering leader, DevEx lead, or delivery transformation role? Use the Role Fit Explorer to map your priorities to the most relevant projects and evidence.

Privacy And Anonymisation

The work here is intentionally presented as anonymised case studies, reference architectures, reusable patterns, and reproducible demos. Some patterns are derived from real consulting and product work, but private product names, customer details, repository details, and proprietary implementation specifics are not exposed.