Loïc Bélec
F1-Grade Architect
I design the nervous systems of high-performance organizations. From factory-to-cloud connectivity to PFLOP-scale compute, I build platforms where performance, reliability, and governance coexist under race-day constraints.
System Metrics
Factory-to-Cloud (ExpressRoute)
Samples per Second (Peak)
HB120rs_v3 Instances
FinOps Optimization
Career Telemetry
A professional timeline of escalating complexity:
From engine metrics to enterprise architecture.
Head of Data Platform
Leading the design and operation of a hybrid data platform, merging on-premise factory systems with Azure to support Formula 1 & endurance programs under strict regulatory constraints.
- Designed a secure hub-and-spoke network with 10 Gbps ExpressRoute, enabling ~3.2 PB/year of data exchange between factory and cloud.
- Structured a centralized lakehouse (Bronze/Silver/Gold) handling 134 daily production pipelines and integrating 11 critical systems (SAP, 3DX, Telemetry).
- Achieved 80% classified data coverage using OpenMetadata (replacing legacy catalogs) & policy-as-code.
- Managed cloud budgets (£1.2M credits) and a €480k professional services engagement, balancing performance with FinOps discipline.
- Integrated HPC capabilities (up to 200 HB120rs_v3 nodes, ~0.9 PFLOP) with storage pipelines for advanced simulation workloads.
High-Performance Data Architect
Architected the high-frequency telemetry platform combining kdb+ and Azure to aggregate wind tunnel, simulator, and trackside data.
- Selected and implemented kdb+ for its superior time-series performance over SQL/NoSQL alternatives.
- Designed end-to-end infrastructure fully automated with Terraform & IaC best practices.
- Integrated Kafka across factories and trackside for low-latency data ingestion (Peak: 20M samples/sec).
Data Engineer
- Consolidated CFD, telemetry and test-bench data into a unified structure.
- Designed object-oriented MATLAB frameworks for race analysis still in use today.
- Automated engine performance reporting, drastically reducing turnaround time.
R&D & Engineering Labs
Pitgun — High-Performance Data Engine
Pitgun is a platform venture combining a live simulation workload, real-time telemetry ingestion and a reusable Rust framework for deterministic data processing. It acts both as a reference workload and as a testbed for building governed, high-performance data systems end to end.
pitgun.dev covers the broader ecosystem and experimental platform rather than the reusable framework itself.
Solar Curtailment Operations Layer
Designing an operational integration layer around EDF Obligation d'Achat workflows to help solar asset operators manage Smart OA constraints, site curtailment requests and billing consequences with better traceability. The project focuses on turning the EDF OA Smart OA curtailment framework into actionable workflows, normalized events and operator-facing automation.
ASK_DETAILSGovernance by Design
Achieved 80% classified data coverage using OpenMetadata & policy-as-code.
FinOps Discipline
Optimized £1.2M cloud credits with ~£150k actual spend via aggressive spot instance usage.
Hybrid Leadership
Bridging the gap between C-level strategy boards and low-level engineering implementation.
Education
Tongji University, Shanghai
Research on U-shaped assembly line balancing (U-ALBP).
Recipient of the CSC Government Excellence Scholarship.
Arts & Métiers ParisTech
One of France’s leading Grande École. Focus on industrial excellence, applied sciences, and mechanical systems.