Product Manager, Decision Engine (up to Principal) in Belfast

Own the decisioning layer behind high-stakes AI. We're hiring a platform Product Manager (up to Principal) to own the Decision Engine in Collective, our sovereign AI operating system for Defence, Government and regulated industries.

Whitespace moves AI from proof-of-concept into real operational use for Defence, Government and regulated industries. Collective, our sovereign AI operating system, connects governed data, domain context, reasoning and action with full auditability. It runs in the cloud, at the edge, on-premise and fully air-gapped, including classified settings. Our applications, including Operational Learning and Decision Intelligence, are live in operational environments today.

The Decision Engine is the set of platform services, components and models that every workflow and application on Collective uses to frame decisions, reason over evidence, generate and test options, and recommend action. Your outcome is decision superiority: better, faster and more defensible decisions in every product built on Collective, including the ones nobody has built yet.

This is a horizontal platform role. You won't own a workflow. You'll find the decision patterns shared across customers, deployments and application teams, then shape them with engineering and our AI and data science team into reusable, assured platform capability.

What you'll do

• Own vision, strategy and roadmap for the Decision Engine as a platform product, from discovery to deprecation.
• Define the decisioning primitives applications compose: objective framing, option generation, optimisation and simulation, reasoning and agent orchestration, uncertainty quantification, explainability and decision provenance, human-in-the-loop controls, and learning from outcomes.
• Decide what belongs in the platform and what belongs in an application. Generalise repeated patterns and say no to bespoke work that doesn't scale.
• Make sharp trade-offs with engineers and data scientists: accuracy vs explainability, reasoning depth vs latency, capability vs deployability at the edge and air-gapped.
• Define how decision capability is evaluated: benchmarks, eval harnesses, red-teaming and production monitoring of model and agent behaviour.
• Own the platform's trust properties: uncertainty surfaced, recommendations traceable to evidence, human authority preserved, an audit trail that holds up years later.
• Treat application PMs and Solutions Engineers as your primary customers, and make the platform their fastest route to decision capability.
• Measure platform outcomes: decision quality, time-to-decision, reuse by application teams, time-to-deploy for new use cases.
• Meet senior Defence, Government and regulated-industry decision-makers to learn how decisions really get made and where they break.

What you'll bring

• A track record shipping complex technical B2B or B2G products into production, ideally platform or API products used by other product and engineering teams. Principal candidates will have owned a platform or product line and its outcomes.
• Platform thinking: turning patterns across many consumers into reusable, well-bounded capabilities, and holding the line on one-offs.
• Real AI depth: LLMs, RAG, agentic orchestration and evaluation, ML and RL, and how they combine with optimisation and simulation to support decisions. You can read an architecture and hold your own with engineers and data scientists.
• A realistic view of non-deterministic systems and how to build in trust, explainability and human oversight where the consequences are high.
• Clear communication from deep technical detail to operational and business value.

Nice to have

• Decision-support products whose output is a recommendation, not a report: constrained optimisation, scenario modelling, forecasting, scoring.
• Defence, Government or regulated decision-making: C2, operational planning, portfolio and investment, risk and compliance.
• MLOps/LLMOps, model evaluation tooling or AI assurance frameworks.
• Delivering AI into on-premise, edge, disconnected or classified environments.
• Active UK Security Clearance (SC), or eligibility to obtain it.

What we offer

• Fully remote for GB-based employees (travel to London or Belfast), or hybrid for Northern Ireland (typically 2 days a week in Belfast).
• Four-day weeks in July, August and December.
• Last Friday of the month off in every other month.
• 3pm finish on all other working Fridays.
• 25 days' annual leave plus bank holidays.
• About 30 extra days back each year on top of your annual leave.
• Shape direction across Collective and our application portfolio at a fast-scaling company.

If you want to build the engine behind high-stakes decisions, where every recommendation has to be trusted, explained and defended, we'd like to hear from you. Whitespace is committed to an inclusive, supportive environment and welcomes candidates from all backgrounds.