siliceum
Field Results

Our field results:
Asobo StudioEuromasterKlafutiMichelinArturia

No empty promises. 5 real projects with results: uptime, latency, performance, scalability.

Asobo StudioMicrosoft Flight Simulator

Flight Simulator 2024 UI Performance Optimization

Push performance to the limit on a Web UI embedded in a C++ game engine, targeting 60fps despite extreme rendering constraints.

PeriodSince 2024
DurationOngoing
RoleEngine Programmers
siliceum Consultants

Context & Missions

Flight Simulator 2024 is an AAA game where every frame counts. The UI (menus, cockpits, HUD) must remain smooth even when the engine pushes the GPU to its limits. No compromise: if the UI lags, immersion is broken.

  • Ensure 60fps on critical screens (menus, cockpits)
  • Low-level profiling: CPU, GPU, memory, cache
  • Optimize the Web UI framework (rendering, data binding)
  • Contribute to fundamental JS framework improvements

Technical Interventions

Performance Diagnostic

  • CPU/GPU/memory profiling with internal tools
  • Rendering bottleneck identification
  • Data binding and UI reactivity analysis

Targeted Optimizations

  • Reduce costly reflows and repaints
  • Optimize observers and data binding
  • Reduce UI memory footprint
  • Lazy loading and asset optimization

Tech Stack

C++Web ComponentsTypeScriptCPU/GPU/Memory ProfilingSIMDMulti-threading

Results

60/30 fps
Critical screens
Menus & Cockpits
JS Framework
Fundamental
improvements
EuromasterMichelin Group

API Architecture & Azure Strategy Europe

Stabilize and optimize a critical European application portfolio serving 17 countries with over 2M requests/day, while adopting an Azure PaaS strategy.

Period2021 - 2023
Duration2 years
RoleAPI Architect & Technical Lead
siliceum Consultants

Context & Missions

Euromaster operates a critical B2B e-commerce platform for 17 European countries, hosted on IBM Cloud VMs. The goal: reliability, performance and controlled costs, with a transition to Azure PaaS for managed services benefits.

  • Modernize Web/API architecture and make it observable
  • Lead the IBM Cloud VM → Azure PaaS transition
  • Establish SRE standards, CI/CD and cross-team practices
  • Define multi-level Web/API QA strategy and load testing

Technical Interventions

Architecture & Observability

  • Modernized and observable Web/API architecture
  • SRE standards and cross-team practices for multiple teams
  • Monitoring, metrics and alerting

Azure Transition & QA

  • IBM Cloud VM → Azure PaaS transition (5 apps, 3 env, 17 countries)
  • Infrastructure as Code and standardized CI/CD
  • Risk-driven, multi-level QA strategy
  • Regression and load testing

Tech Stack

Spring BootAngularAzure PaaSTerraformAzure MonitorAPMK6GitLab CI

Results

99.9%
Uptime
>2M req/day
P99 < 200ms
Latency
critical endpoints
KlafutiSaaS Startup

Production-Grade SaaS & Long-Term Support

Build a production-ready B2B accounting SaaS from V1, then support growth over 2 years with a solid technical team.

PeriodSeptember 2023 - December 2025
Duration2 years
RoleCTO & Architect
siliceum Consultants

Context & Missions

B2B SaaS startup developing an accounting process automation solution. Complete support: from production-grade V1 to team and feature scale-up.

  • Lay technical foundations: architecture, CI/CD, reliability practices
  • Build Cloud infrastructure (Azure, Terraform, GitLab CI)
  • Create Angular Design System and product visual identity
  • Develop core features (Java Spring Boot & Angular)
  • Migrate to Clever Cloud for simplified operations
  • Support scaling and product evolution over 2 years

Technical Interventions

Reliability from Day 1

  • Scalable architecture adapted to opportunities
  • CI/CD with Blue/Green deployments and Feature Flags
  • Multi-level automated testing
  • Observability and alerting from launch

Clever Cloud Migration

  • Mapping and migration plan Azure → Clever Cloud
  • CI/CD adaptation for PaaS deployment
  • Controlled switchover with zero downtime
  • Simplified operations and controlled costs

Tech Stack

JavaSpring BootAngularTypeScriptClever CloudTerraformK6

Results

Feedback-driven
Product roadmap
driven by the field
Production-grade
SaaS from V1
Scalable architecture
2 years
Support
Growth & evolutions
MichelinIoT & Connected Fleets

Real-time IoT Platform for Truck Fleets

Collect, process and analyze in real-time data from thousands of IoT sensors embedded on trucks across Europe.

Period2019 - 2022
Duration3 years
RoleTech Lead & Backend Developers
siliceum Consultants

Context & Missions

Michelin equips professional fleets with sensors that continuously measure tire pressure and wear. This data must be collected, processed and delivered to fleet managers in real-time to optimize maintenance and safety.

  • Design event-driven microservices architecture
  • Integrate IoT sensors with real-time data collection
  • Build data pipeline (Azure Event Hub, Service Bus)
  • Deploy on Kubernetes with automatic scaling
  • Industrialize the solution for European fleets

Technical Interventions

Real-time Architecture

  • Java 11/17 microservices with reactive Vert.x
  • Event-driven architecture for high availability
  • Azure Event Hub for massive ingestion
  • Load spike management (thousands of sensors)

Industrialization

  • Azure Kubernetes (AKS) multi-environment
  • Automated GitLab CI/CD pipelines
  • Grafana & Splunk monitoring
  • Progressive multi-country rollout

Tech Stack

JavaVert.xSpring BootAzure Event HubService BusKubernetesDockerMongoDBInfluxDBGrafanaGitLab CI

Results

Europe
Deployment
Truck fleets
Real-time
IoT data
Tire pressure & wear
Scalable
Architecture
Thousands of vehicles
ArturiaMusical Instruments

Audio Synthesizer Performance Optimization

Reduce latency and CPU usage of software synthesizers to ensure a smooth experience for professional musicians.

Period2023
Duration6 months
RoleC++ Performance Expert
siliceum Consultants

Context & Missions

Arturia creates software synthesizers used in studios and live performances. For a musician, every millisecond of latency is noticeable. The challenge: optimize real-time audio processing while maintaining a responsive UI.

  • Profile the audio engine and identify bottlenecks
  • Optimize DSP processing and CPU consumption
  • Improve OpenGL UI rendering
  • Reduce audio latency to imperceptible levels

Technical Interventions

Low-level Diagnostic

  • CPU profiling with Tracy Profiler
  • DSP hotspot identification
  • Cache miss and branch misprediction analysis
  • Benchmarking on real user scenarios

Real-time Optimizations

  • SIMD vectorization of signal processing
  • Elimination of real-time allocations
  • UI rendering pipeline optimization
  • Overall audio latency reduction

Tech Stack

C++JUCEOpenGLTracy ProfilerSIMDDSPReal-time Audio

Results

Major gains
CPU Performance
Tracy profiling
Minimal latency
Real-time audio
Imperceptible
Smooth UI
OpenGL Rendering
Optimized

Your project deserves
the same results

Free discovery call. We discuss your context, your challenges, and see if we can help. Response within 48h.