A personal career operating system: jobs, research, funding, people and events
One pipeline runs every day: it follows startup funding and company news, research positions and networking events, scores how well each role fits my profile, and files the leads I share from my phone. It started from santifer's open-source career-ops; on top I built the daily run, the sources beyond job boards, the scoring and the web dashboard.
Built with AI coding agents; I set the design, the rules and the tests they have to pass.
What it covers
Where it looks
- Opportunities
- Company job boards through ten hiring systems' own interfaces, the Federal Employment Agency's public API and venture-portfolio boards.
- Companies
- Each company's boards, checked: a board naming another employer is turned down, an empty one flagged.
- Funding and news
- A weekly digest: a robotics or autonomy startup that raises money enters the database before it hires.
- Research
- Doctoral positions in their own tracker, on a separate scorecard.
- Events
- Hackathons, meetups, fairs and founder programmes, checked at the organiser's page.
What I bring to it
- Leads
- Roles, doctoral positions, events or companies I spot on my phone, each filed for evaluation.
- Capture
- Application Tracker, my browser extension, saves postings to Notion; Career-Ops reads the statuses back.
How it judges
- Evaluation
- A model grades each requirement of a role; code turns the grades into a score, with hard caps.
- One dashboard
- Today, jobs, leads, people, research, events, news and companies in one local window.
- Guards
- Paid data has a weekly cap; unattended runs are told to treat page text as data, never as instructions.
What is santifer's, and what is mine
Built on santifer's open-source career-ops (MIT), which already evaluated jobs and scanned three hiring systems. I made it run every day on its own and widened it into everything above.
From santifer/career-ops, MIT
- Evaluating a job
- A scanner for Greenhouse, Ashby and Lever
- Evaluations in batches
- A terminal dashboard
Mine
- The daily run, on a schedule
- Seven more hiring systems
- Public APIs, venture boards and news
- Finding companies' boards and checking their health
- Computed scores with hard caps
- Leads, research and events trackers
- The web dashboard
The test suite grew from about 300 lines to about 5,800, run against temporary folders and never against real data.
Next in the Builder Lab
MechCopilot