BETA CAE Systems

BETA CAE Systems builds simulation software used by automotive, aerospace, and energy companies to analyze structural integrity, crashworthiness, and computational fluid dynamics. The software handles geometric models with millions of elements and runs analyses that take hours or days on HPC clusters. The GUI is the engineer's primary interface to all this complexity — every feature engineers use ships through it. Performance, stability, and cross-platform behavior are non-negotiable: a UI bug in production can invalidate a multi-day simulation run.

For 10 years I built the desktop interfaces that mechanical engineers used every day to do their actual work — navigating geometric models with millions of elements (car bodies, aircraft wings, energy infrastructure), configuring physics simulations that ran for hours or days on HPC clusters, reviewing analysis results that informed billion-dollar engineering decisions.

These interfaces shipped to tier-1 automotive, aerospace, and energy customers worldwide. The software runs in production for decades — meaning UI bugs aren't cosmetic, they invalidate multi-day simulation runs and cost real engineering hours. **The deep Qt expertise built over a decade in this environment is rare** in today's web-first engineering market.

**Beyond the desktop — real-time engineering collaboration**

I also helped build a web application that let mechanical engineers in different offices — or different companies on joint projects — share live simulation sessions and remotely operate the desktop tools across the network. Real-time collaboration for technical power-users in the early 2010s, **years before Figma made multiplayer canvases mainstream (2016)** and before modern AI co-pilots existed.

The architectural problems we solved (synchronizing distributed state across high-latency networks, low-latency peer connections through corporate firewalls, remote control of complex native applications) are the exact same problems that show up today in AI agent collaboration tools and multiplayer AI dashboards.

During my decade there, the engineering organization grew from ~100 to 250+ people, and the products matured into the standard tools for many tier-1 automotive and aerospace customers worldwide. The company was acquired by Cadence Design Systems in 2024 (transaction valued at $1.24B — 5 years after my engineering tenure ended).

**What this work demonstrates:** • How to build software that runs in production for decades without rotting • How to design UIs for technical users who need power, not consumer-grade simplicity • How to operate inside fast-growing engineering organizations where code lifetime matters • How to handle performance-critical rendering and large-dataset visualization • How to build real-time collaboration systems before "multiplayer" became a frontend trend • How to bridge complex native applications with web interfaces (a pattern increasingly relevant for AI agents)

The UI patterns that work for engineers managing complex simulations are the same patterns that work for operators managing AI agents. Modern AI infrastructure has rediscovered problems we were already solving a decade ago — the difference is that now the "other participant" on the call is often an AI agent.