Framing the problem.
State the hypothesis.
Are we all seeing the same thing.
Do we have empathy for our customers?
Who is the real customer?
What Job to be Dones do they have?
I turn complexity and AI Output into clear, state-of-the-art software. For founders and product teams that need senior product judgment, interaction design. I deliver testable pilots or functional product. I work in strategic setups or fully operational. 10+ years experience.
My Approach
I help companies reduce the noise and increase the signal.
I create value through intelligence, creativity, aesthetics and flipping the switch from zero to one.
I have worked in different setups, in different roles. I challenge, I collaborate, ask questions and get things done.
A selected list of clients over the years.
Compress your product loop. Synthesize where it creates value.
Synthetic methods accelerate exploration and implementation. Real behavior determines what deserves to scale.
Synthetic Design Methods
Synthetic Coding
Synthetic Analysis
Framing the problem is crucial decision making. Evidencing gives us a credible thesis. Always looking for the shortest path from zero to one.
Vibe Coding is great for simple tech problems. Software architecture is where things really start to come together.
A selection of
Framing the problem is crucial decision making. Evidencing gives us a credible thesis. Always looking for the shortest path from zero to one.
Strategic UX digs deeper into the problem.
Successfully leading the most business critical inhousing project, while managing stakeholders and interviews across departments.
Built the foundation for the most central dashboard in the corporation.
Head of Products
Mitch Seguin
Sahil Gulhane
Product Management
Somto Zikora
Open Source
Agentic work, is the new normal. Cutting down on the agentic slack too. Too much signal is still noise.
I have been building my own tools. AI is only as intelligent as the pipeline. Product outcomes are not guaranteed, when an agent doesn't know what to do, where to look or what the right context is for the task.
I built my own package to code faster, save on tokens and keep most of the workflow deterministic instead of letting the agents handle the fuzzy work.
It doesn't matter where and how the LLM is hosted. The tools stay the same.