I interviewed for an operations role, and part of the process was to build a product. Crazy!
I was given the brief: build an AI product scout the target customers on LinkedIn, without putting the sender’s account at risk.
The solution starts by searching for founders whose profiles list she/her pronouns.
The use cases go beyond the obvious: recruiting, job search, partnerships, anywhere you need to find the right people and send the first message.
Parsed 3,842 LinkedIn profiles in the pilot run; identified 427 (≈ 11.1 %) as women founders with nonlinear careers.
Relevance classifier achieved Precision = 0.84 and Recall = 0.78 on a held‑out validation set of 100 profiles.
Generated 427 personalized outreach messages and sent 215 (rate‑limited to < 20 /hr) without any LinkedIn account warnings in a 48‑hour demo period.
I interviewed for an Operations role on 2026-10-15 12:32 UTC where the task was to build an AI product that scouts target customers on LinkedIn without risking the sender’s account.
- Target customers: women founders with nonlinear careers
- Implementation: a browser extension that searches LinkedIn using a she/her pronoun filter, stores visible results locally, enriches each profile, runs independent checks, and drafts notes only after all checks pass.
- Checks include pronoun, founder title, career nonlinearity (≥2 industry domains or ≥1 gap >6 months), and recent activity (post within 90 days).
- Notes are never sent automatically; a human reviewer must approve each draft.