I'm a data & AI platform leader. When I started looking for my next role, I did what I'd do at work: built a system. Sourced roles programmatically, scored them for fit, tracked everything, and mined my network — then documented the whole thing in the open.
Every section below flips to reveal the code behind it — the whole search, engineered in the open.
A job search is a data problem: too many listings, weak signal, and no easy way to compare fit. I treated it like building a platform — source, normalize, score, track, act.
Pull roles from a live jobs data lake (1M+ listings) and Indeed, filtered to my level, geography, and comp floor.
A transparent rules engine rates every role 0–100 for fit against my background, weighting strengths and flagging gaps in plain language.
Everything lands in a colour-coded tracker — status, source, comp, and the fit rationale — so nothing slips.
Tailored résumés and cover letters per target, applications driven to the review step, and network mined for warm intros.
Live view of the pipeline. In public mode, companies collapse to their industry and compensation is hidden — the shape of the search stays, the sensitive detail doesn't.
| Company | Role | Comp | Status | Fit |
|---|
Instead of scrolling job boards, I query them. Roles come back structured, get de-duped against what I've already seen, and are ranked before I ever read one.
Fans out searches across multiple angles — by title, by seniority, by stack — so one blind spot doesn't hide a good role. Filters to Director/VP/Principal level, remote or DC-metro, and a comp floor, then sorts by freshness and fit.
Recruiters optimize for keywords; candidates should optimize for fit. Scoring every role the same way removes wishful thinking and surfaces the stretch-but-real opportunities I'd have scrolled past.
The same discipline I bring to building data platforms, pointed at my own career.
The best roles come through people. I take an official export of my network, group it by employer, and cross-reference against companies that are actively hiring for roles that fit — then reach out where there's a warm path.
Use the platform's own data-export (ToS-friendly) — names, employers, titles, connection dates.
Roll connections up by employer and tag each with industry and seniority.
Join against live openings — surface contacts who work somewhere with a role that fits.
Auto-draft short, personalized referral asks — ranked by the contact's seniority and relevance.
Same career, sharper story. I was titled in marketing but doing platform engineering — so I repositioned to lead with what I actually build, and tailored a version for each kind of role.
Dropped the marketing framing; lead with platform, AI, and architecture.
Simplified to Senior Director — function over department.
Opens on the unified data platform and production AI, not campaigns.
Three tailored résumés — commercial data/AI, cleared solutions architect, and federal — plus per-role cover letters.