Start with the problem
I want to understand the workflow before deciding what to build. Where does time go? What needs judgment? What would make the work easier? Those questions guide the product and the technology.
Build, test, learn
I prefer a small working version that teaches us something. Early feedback helps me decide what to keep, what to change, and what to stop building.
Stay close to the details
Product decisions and technical decisions shape each other. I work through the constraints, examine the tradeoffs, and take responsibility for how the pieces fit together.
Make AI useful in context
Model capability is one part of the system. Data, evaluation, interface design, and human judgment all matter. I care about whether the application helps someone complete real work.
Work with uncertainty
Complete information rarely arrives. I make assumptions explicit, test the ones that matter most, and adjust as I learn.
Keep communication clear
Clear writing helps people think together. A short conversation can resolve ambiguity, and a written decision helps everyone move forward with the same context.
Make feedback part of the work
I value direct feedback while there is still time to act on it. I want to know where the product or process is creating friction so we can improve it.
Build systems that help
Automation, dashboards, and playbooks should make work easier. I look for repeated problems and build systems that give people more room to focus on the difficult parts.
Keep learning
I ask questions, admit gaps, and learn what the problem requires. Curiosity keeps me moving when the answer is unclear.