Digital Risk Platform – Multi-Sport Pricing & Exposure System
What I did: Took the company's first in-house risk platform from vision to production, and built the first cross-functional product team (10+ engineers, data and design) to deliver it.
Result: Replaced a third-party provider, saving $4m a year. Bringing AI-assisted development into the team lifted front-end delivery by 20%.
Summary
A full modernisation of Angstrom’s risk architecture, unifying pricing, exposure, liquidity, and trader controls into one cross-sport platform used across NBA, NFL, and MLB.
Context
Angstrom operated multiple fragmented tools across sports. Traders lacked consistent controls, visibility, and reliable workflows, creating operational bottlenecks during high-pressure events.
Problem
Pricing systems were fast, but risk controls were siloed. Teams needed unified visibility, actionable guardrails, and automation to reduce manual workload and improve reaction speed.
Constraints & Stakes
- High-risk, real-money environment where seconds matter
- Cross-team dependencies: FE, BE, modelling, trading, data
- Hard seasonal deadlines (NBA/NFL kick-offs)
- Complex distributed architecture (Kafka streams, microservices)
My Role
As Product Owner for the entire digital risk platform, I owned roadmap definition, cross-sport alignment, prototyping, delivery leadership, and stakeholder communication across Entain, BetMGM, and internal teams.
Actions
- Discovery: Interviewed 30+ traders to map workflows & pain points.
- Architecture: Designed a unified risk schema across exposure, limits, automation, and pricing impact.
- Design: Introduced AI-assisted development and Figma → React prototyping, increasing front-end delivery by 20%.
- Team formation: Built the first cross-functional Risk Product Team.
- KPI Framework: Defined adoption, latency, exposure accuracy & automation metrics.
- Rollout: Coordinated multi-market onboarding across NBA, NFL & MLB trading teams.
Impact
- 20% increase in front-end delivery through AI-assisted development.
- Unified risk behaviour across all major US sports.
- Created the foundation for a proprietary feature and future automation pillars.
Where this applies
The problem underneath is common to any business that takes on risk in real time: see exposure as it builds, act on it within agreed limits, and be able to show what happened afterwards.
- Real-time exposure and limits: trading desks, payments and fraud, lending, insurance accumulation
- Automated decisions with human override: pricing engines, credit decisioning, dynamic pricing in travel and e-commerce
- Audit trails and tiered permissions: any regulated product, from financial services to healthcare
- Replacing a vendor with in-house capability: build-versus-buy calls in any platform team