← All Work
Operating Infrastructure · Institutional Memory

How do you preserve good judgment after the people who developed it leave?

Hiring Compass is my answer — a system built to hold what great hiring managers already know, so it doesn't disappear when they do.

1,700+Candidates Processed
25Resumes Per Batch
6+PE Profiles Built
01 — The Question

What problem did I notice?

Thirteen years of watching the same thing happen: a hiring manager builds real judgment about what "strong" looks like — then a search ends, a team changes, or a new partner comes on, and that judgment disappears. The next search starts from zero, rebuilding knowledge that already existed somewhere, in a conversation nobody wrote down.

The knowledge was never the problem. No system was built to hold it.

Ambiguity by default. "I'll know it when I see it" isn't a rubric.
Inconsistent evaluation. The same candidate, read differently by different partners.
No infrastructure for volume. Speed and quality in constant tension, with no system to resolve it.
02 — The Decision

What did I choose to believe?

The insight didn't come from a product brief. It came from years of proximity to executive-level hiring decisions — watching what hiring managers noticed first, what made them hesitate, what made them move fast. That pattern recognition was the missing infrastructure.

"The tool wasn't built because I wanted to build AI. It was built because I kept seeing ambiguity repeated across hiring managers and wanted to make it teachable."

I decided institutional knowledge should compound instead of disappearing — and that a system could hold it without replacing the judgment that created it. Three beliefs shaped everything that followed:

I
Human judgment is always final. Scores are alignment suggestions, not decisions.
II
Every search should make the next one smarter. The system accumulates. It doesn't reset.
III
AI should reduce cognitive load, not replace expertise. It handles volume and pattern recognition. The human handles context and relationships.
03 — The System

How did I embody that belief?

This is easier to show than to explain:

Sample Output
Hiring Compass · Scoring Output
Web Experiences L5 — Candidate Review
★ Strong Yes
D1
Interactive & Web Craft
3/3
D2
Brand Expression & POV
3/3
D3
Design Systems & CMS
2/3
D4
Motion & Interactive Range
2/3
D5
AI Fluency (Claude Code / Cursor in production)
3/3
D6
Taste, Restraint & POV
3/3
Illustrative output. Scores are alignment suggestions — human review is always determinative.
Session Opener — Branching Flow
Activate Hiring Compass
What do you need today?
Mode 1
Score Candidates
Resume batch → D1–D6 scoring → Strong Yes / Maybe / No
Mode 2
Update Profile
New PE signals → rubric update → write-back confirmed
Mode 3
Sourcing Brief
Role context → target companies → search strategy
Mode 4
Calibration Review
PE feedback → rubric refinement → pattern synthesis
Write-back rule: After every session, results are written to skill files before closing. Institutional knowledge never disappears.

Underneath both: every hiring leader's profile holds their actual rubric and calibration history — not the job description. Every session writes back before it closes, so nothing learned in one search has to be rediscovered in the next. And a strong candidate who isn't right for one role can surface for another, but only after a real fit check, never as a shortcut.

The first version lost context mid-session and had to be rebuilt — which is exactly why the structure holds up the way it does now. The system got better because it kept failing in public, not because I designed it correctly the first time. This was built solo — every architectural, aesthetic, and philosophical decision.

04 — The Result

What changed?

1,700+
Candidates scored in early use across active PE searches
25
Resumes reviewed in 8 minutes — vs. 3–4 hours manually for the same volume
6+
PE profiles built, each capturing rubrics, hard filters, and calibration history
"I see her as a peer and thought partner and find that she's constantly helping me stay on track and reminds me of what matters when reviewing candidates... Very AI fluent, and leverages new tools to make the recruiting function even more effective."
Design Leader — Head of Design
"Building Claude skills for application reviews with Bri helped us create a shared brain so much faster. We were able to externalize and align on our candidate criteria, and get calibrated on the role with lightning speed. It made the application review process so much smoother. The shared skills also helped me avoid bias and inconsistency in reviewing applications – it reminded me of the skills and experience we'd previously decided was most important for the role."
Alaine Mackenzie, VP of Design for Core Experiences and AI
Epilogue

What Changed in Me.

Building Hiring Compass didn't just change how I recruit. It changed how I understand my own career.

I stopped thinking of myself as a recruiter who uses tools and started thinking of myself as a designer who builds them.

I realized that the work I've always done — translating ambiguity into structure, earning trust to access real problems, building systems that outlast me — traces back to a discipline: service design. What I actually build, I call operating environments.

I understood that institutional knowledge isn't just a recruiting problem. It's an organizational design problem. And I know how to solve it.

I learned that AI doesn't replace what I do. It finally gave me the medium to build what I've always imagined.

I stopped waiting for someone to give me the title that matched the work I was already doing.

"I don't build systems because I love systems.
I build systems because I care about the people who have to use them."
Scores and outputs generated by Hiring Compass are alignment suggestions based on resume-level signal only and are not final hiring determinations. Human review is always determinative. — Bri Majors

Editorial Note: Names and identifying details have been changed to protect confidentiality. The work itself has not.