The Brief
Context
For each case below, answer the following 4 questions:
What is the core problem that actually needs to be solved?
Without considering AI first, how would you design a better non-AI workflow/process?
Where would you inject AI into that workflow?
(Feel free to elaborate: which parts should rely on deterministic rules, existing domain knowledge, or require custom evaluation/prompting for the AI?)
How would you evaluate and continuously improve this system over time?
The Task
Case 1 — Candidate Fit
Currently, evaluating whether a candidate is a good fit for F. relies heavily on the individual HRM or SME conducting the review.
For the exact same candidate, one reviewer might rate them very highly while another finds them unqualified. Criteria such as problem-solving, ownership, communication, or "cultural fit with F." are interpreted differently across reviewers.
The team wants to reduce subjectivity and personal bias, without turning hiring into a rigid, robotic scoring rubric.
How would you design a solution for this?
Case 2 — Content Reviewer
The content team has internal methodologies, frameworks, and quality standards for evaluating written pieces.
Currently, reviewers have to juggle and memorize a large set of rules during reviews. Consequently, feedback consistency varies widely across reviewers, certain edge cases or errors slip through, and comments sometimes fail to pinpoint root causes.
The team wants an AI Reviewer to assist the review process.
How would you design a solution for this?
What you need to submit
How would you design a solution for this?
Tip: If you are unsure how to structure your answers, follow this straightforward chain:
Problem → Better workflow → AI fits where? → How it gets better over time
Reflection (End of submission)
Briefly share:
Which parts of your solution were assisted by AI?
Were there any suggestions provided by AI that you decided not to use, and why?
Submit Your Challenge