I created the open-source connectivity software that runs in billions of IoT devices. I’ve shipped more than 30 connected products with clients including ABB, Electrolux, Husqvarna, tado, and LiFX. For 12 years I ran Thingsquare as CEO. Now I focus on two things: teaching teams how AI changes IoT product testing, and helping companies build connected products using agentic development.
Hands-on Course: AI-Powered IoT Testing at Scale
Testing connected products at scale is brutal. Hardware permutations, firmware variants, field conditions you can’t reproduce in the lab. Most teams test manually until something slips through to production.
AI agents change this. They can generate test scenarios, simulate device fleets, find edge cases humans miss, and run regression suites that would take a team of ten. I’ve built these systems for real products and I teach your team how to do the same.
Hands-on courses for engineering teams:
- How to use LLM agents for automated test generation across firmware variants
- Building AI-driven device fleet simulation
- Scaling test coverage without scaling headcount
- Integrating agentic testing into your existing CI/CD pipeline
These are working sessions. Your team leaves with infrastructure they can run on Monday, built around your actual product.
Project Engagement: AI-Accelerated IoT Product Development
Building connected products is hard. Building them with a small team under real constraints is harder. Agentic AI makes it possible to move at a pace that was previously out of reach. But only if you know where it works and where it doesn’t.
I’ve spent decades shipping IoT products and the past two years figuring out exactly where AI agents fit in the development workflow. Firmware, cloud backends, mobile apps, protocol implementations. I know which parts an agent handles well and which parts still need an experienced engineer making the calls.
Hands-on engagements:
- I embed with your team and work alongside your engineers
- We set up agentic workflows tailored to your product and stack
- I bring the IoT domain expertise; we use AI to multiply your team’s output
- Typical engagement: 3 to 6 months
I make technical decisions, review architectures, and get the product out the door.
Frequently Asked Questions
How is AI used for IoT product testing?
AI agents generate test scenarios from firmware specifications, simulate fleets of hundreds or thousands of devices, and find edge cases that manual testing misses. They run continuously against new firmware builds and catch regressions before production. The result is broader coverage with a smaller team.
Who are these courses for?
Engineering teams building connected products who already have a product in development or production. You don’t need prior AI experience. You do need firmware engineers who want to ship faster with fewer escaped bugs.
What does a typical product development engagement look like?
I join your team for 3 to 6 months, typically 2 to 3 days per week. We identify which parts of your development workflow benefit from agentic AI, set up the tooling, and I make architecture decisions alongside your engineers until the team runs independently.
Get in Touch
I take on a small number of engagements and would love to hear what your current challenges are. Send me an email at adam@dunkels.com.
