Every Camundi has a before and an after with AI. Before AI became part of how we work. And after it started changing the way we think, build, and solve problems.
This series is about that journey. The moment someone tried something, broke something, improved something, or built something they did not expect.
Each month, one Camundi from somewhere across the business walks us through their AI adoption journey. An open look at how people are figuring it out as they go, across teams.
If you have been curious but have not started yet, this is for you. If you have started and feel like you are doing it wrong, this is for you too.
How one engineer used AI-first development to rethink engineering velocity
Who are you and what do you do at Camunda?
I’m Josh Wulf, the technical lead of the cybernetic engineering team. We unlock and unblock our people to leverage their own and others’ strengths to build cybernetic systems that raise the reliability of the human-LLM cyborg across the spectrum, unlocking safe velocity.
What did the work look like before and why did you decide to try AI on it?
I was the Camunda 8 Node.js SDK maintainer. I created the SDK in the community as an open source project and brought it to Camunda when I joined seven years ago. A year ago, I was bored of manually implementing things that could be automated but felt I didn’t have the capacity to reach across our tech stack and automate them. I was also frustrated by some classes of defects popping up in our tech stack repeatedly, signaling unaddressed defect-generating surfaces. My day-to-day work and the scale of the change that would be needed kept me from addressing it.
A friend of mind, a CTO at an AI company, sent me a simple message in June last year: Our developers are getting great results from the latest OpenAPI model.
I pointed Copilot at CPT (Camunda Process Test) and told it: “Reimplement this in Node.js.” The result was far from perfect but also far exceeded anything that I thought was possible. When I saw that, the sky opened up, and I had a vision of the future.
What did you actually build or change and what’s the impact/result?
In July 2025, I was maintaining just the Node.js SDK for Camunda 8. As of June 2026, I’ve published three additional Camunda 8 SDKs, extending our market to two new programming languages. I implemented an industry-leading innovative dominant type system in these new SDKs and started academic partnerships with two New Zealand universities to extend our strategy into automated API test generation, exploring the implications of research I published in 2011 while at Red Hat.
I pioneered the Dark Factory pattern at Camunda and the systems and processes that it takes to let LLMs stream features into a code base at high speed with safety, calling for the attention of a human when it is needed.
I was able to address several surfaces across the stack to enable the generation work and increase the system reliability end to end, collaborating with both core features engineering and QA.
At the company kickoff in January of this year, I was hype man for Volker when he created `c8ctl`—an innovative and powerful CLI for Camunda 8—building on our new JS SDK, and then took over as core maintainer in April. I’ve been able to train a university intern, Emily, who has since joined the company full-time in AI-first development.
With Volker and Emily, this year we started to explore how people work together, leveraging each others’ strengths to understand and replicate cybernetic systems that provide safety at the incredible velocity that the developer-LLM fusion unlocks. Volker took the role of product designer leaning fully into LLM velocity and blazing a bold new trail. I took the role of code architect, following in his footsteps and refactoring the code base and supporting structures for safety. Emily and others took the role of developer-LLM cyborgs, landing innovation at a fierce rate and challenging the cybernetic system of the `c8ctl` Dark Factory to absorb them safely and quickly.
What went wrong, and where does it still let you down?
There is nothing wrong; and the next frontier is building a further cybernetic system—the fusion of people, processes, and feedback loops—that empowers our people to build the cybernetic systems of the Dark Factory pattern around their code bases, unlocking innovation engineering at scale and speed.
If you were starting from scratch tomorrow, what would you do differently?
I am now the product of one year of LLM-first development. If I were ported back one year knowing what I know now technically, I would have a completely different strategic vision. This is an interesting question—if I put myself one year in the future from now and look back, that changes my perspective in the present. There are two futures: one where LLMs go away, and we maintain the code we are now generating by hand. In the other future, we continue to accelerate and build the world’s foremost agentic engineering organization in the area of process automation. In that future, people refer to “the Camunda way of engineering” as a standard to aspire to. This is the future that I’m committed to.
What can other teams steal or try right now?
The key concern developers have when an LLM can generate features and solutions faster than they can read or reason about them is, “How can I know that this is correct?”
That question is worth sitting with seriously, like a Zen koan, because it reveals the key to scaling. A system that proves the correctness of LLM output is key, and it isn’t reasoning about the output at human speed. That approach constrains the velocity if you do it and reduces safety if you don’t and have no other answer. Every scenario is distinct, but human innovation can generate a solution to proving correctness at scale. Do it enough times and principles start to emerge. In the post-LLM world, this is one of the areas where human innovation is critical to unlock value.
Come build with us
The problems Camunda is solving are hard. The people solving them are interesting, lifelong learners, who bring curiosity, empathy, and persistence to everything they do. The work we do at Camunda isn’t for everyone, but if this kind of thinking energizes you, you’ll likely feel right at home.
If this series made you think, "I would love to do that,” you probably would. Come see what we're building. Explore our open roles here.
