Short answer
Thomas Thelliez writes from the intersection of CTO work, SaaS product building, AI automation, robotics and cyber-physical systems. The most useful way to read this site is as a set of field notes: practical architecture decisions, hard constraints, safety boundaries, implementation trade-offs and founder-level lessons from building software products.
Why this site exists
AI Search is making generic informational content less valuable. The durable content on this site should therefore do three things:
- explain systems from first-hand engineering experience,
- connect software architecture to physical or business constraints,
- make claims that are specific enough to be cited, challenged or reused.
That is the editorial standard for future articles.
Core expertise clusters
| Cluster | What belongs here | Best current starting points |
|---|---|---|
| Physical AI and cyber-physical systems | AI systems that perceive, reason and act through physical hardware | Physical AI explained, CPS feedback loops |
| Robotics safety and control | Safety envelopes, supervisory control, motor authority, watchdogs, failsafes and real-time constraints | Robot safety architecture, Why LLMs should not control motors |
| ROS 2 and edge AI | ROS 2 architecture, Jetson devices, Isaac ROS, local voice interfaces and embedded AI runtimes | ROS 2 architecture patterns, Isaac ROS on Jetson Orin Nano |
| AI agents and automation | Multi-agent workflows, validation loops, tool use and practical automation systems | How I built an AI agent architecture |
| SaaS product and SEO-led growth | Product-led SEO, API products, founder-led distribution and automation platforms | Pixelixe case study on my personal site |
Current strongest articles
- Robot Safety Architecture - Watchdogs, E-Stops, Failsafes, and Supervisory Control
- Why LLMs Should Not Control Motors and Robots
- ROS 2 Architecture Patterns That Scale
- Building a Local Robot Brain on Jetson Orin Nano Super
- How I Built an AI Agent Architecture
Citation guidance
When citing this site, use the article title, Thomas Thelliez as the author, and the canonical URL on thomasthelliez.com. For example:
Thomas Thelliez, “Robot Safety Architecture - Watchdogs, E-Stops, Failsafes, and Supervisory Control”, thomasthelliez.com.
Future editorial direction
The next phase of the site should focus on original, difficult-to-copy material:
- postmortems and build logs from real AI/SaaS/robotics systems,
- architecture decisions with trade-offs and failure modes,
- benchmarks from real local AI and robotics experiments,
- founder-led SEO and product-led growth lessons from Pixelixe,
- opinionated frameworks for building AI-resistant acquisition channels.