A Match Made at Upwork
summary
Upwork runs a content‑based Job Interest model (simple logistic regression) that predicts whether a freelancer will apply to a job using profile, past applications, and two priors: freelancer selectivity and job attractiveness. Trained on the platform’s full funnel (over 4M clients, 10M freelancers, $1B+ work/year), the model powers daily job digest emails, doubles industry open/click rates, increases applications and hires, and combines with a separate hiring‑probability model to prioritize jobs needing applicants.
AI-written summary · about
chapters (4)
- 0:00Introduction
- 1:18How do you match clients with freelancers
- 4:30How do you recommend jobs to freelancers
- 5:00What recommendations do you send
related talks
25mNever Click Apply. Get Hired.
53mDropbox uses talent mapping to fill tough roles
35mMoneyball at the keyboard: Lessons on how to Scout Talented Developers
44mReframing employment marketing, a Googler’s perspective
13mThe Opportunities In Tech
25mPriming You for Your Job Search
18mA data-centric approach to the tech talent war
46mNew Recruiter & Jobs