An open-source RPA system that automates cover-letter submission on BOSS Zhipin.
01 > SCAN MATCHING JOBS
02 > ANALYZE DESCRIPTION
03 > DRAFT REPLY WITH LLM
04 > QUEUE BROWSER ACTION
CLIENT OR PRODUCT
Open-source automation project for BOSS Zhipin job applications.
MY ROLE
Automation engineer / open-source contributor
THE PROBLEM
Job seekers were stuck doing robot work: open listings, tailor materials, click through forms, and repeat.
CONSTRAINTS
- •Third-party web UI that can change without notice
- •Need for LLM-assisted content generation
- •Reliability over demo-only scripts
- •Open-source usability for other operators
THE INVESTIGATION
Broke the application flow into stable automation steps and identified where LLM assistance actually reduces work versus where brittle selectors fail.
THE STRATEGY
Combine browser RPA with LLM-assisted cover-letter generation so the system handles both navigation and content drafting — not just one half of the chore.
WHAT I BUILT
- •RPA workflow for BOSS Zhipin submissions
- •LLM-assisted letter generation path
- •Open-source project packaging for reuse
- •Operator-oriented automation design
THE HARD PART
Keeping automation useful when the target site and job-market context both shift — designing for recovery and iteration, not a single happy path.
THE VERDICT
- •Public open-source automation available for others to run and extend
- •End-to-end path from listing to submitted materials
- •Concrete example of AI + RPA in an operational workflow
RELATED CASES
© 2026 Sidi Li. All rights reserved. | Built with ❤️ and determination. | Thanks for visiting!