← Work
Founder & Product · Inzpira · 20222024

SkillWee

A custom evaluation engine that scored decision quality, not memory recall. Built before the term 'agentic eval' was on everyone's lips.

LLMSimulationCareerEvaluation

SkillWee dropped learners into realistic career scenarios. A product launch in trouble, a hiring decision under constraint, a conflict between two team leads. It measured how they reasoned, not what they remembered. The evaluation engine scored along axes the learner could see: completeness, structure, trade-off awareness, communication.

1,000+ users at peak. The takeaway that fed forward into ATOM was simple. The most useful AI products don't replace judgment. They make judgment legible and improvable.

Impact

  • 1,000+ active users at peak
  • Custom rubric-based eval engine