About this video
This live StellarPeers session teaches you to treat risk as a design input, not an afterthought. That mindset runs across framing, prioritization, solutions, and metrics.
First, it covers the rubric interviewers use to score an answer, including problem framing, AI risk awareness, product judgment, solution thinking, metrics depth, and intellectual honesty.
Then, it walks through a nine-step framework for the question, “How would you design a teen safety experience for ChatGPT?”
Next, Malena Mesarina puts that framework into practice live with the group. She prioritizes drug-related curiosity as the harm worth designing for, then names the risk of each solution before choosing a classifier-based MVP.
Finally, she deduces metrics from goal achievement, mechanism health, and ecosystem signals, coaching candidates in real time.
Overall, the session shows that refusing to help teens is not automatically safe, and that real product judgment means being honest about what your metrics cannot tell you.