Materials · synthesis/02-board-ranking.md

# Board ranking (orchestrator synthesis)

Sources: the five scout reports in `screening/`. Ranking and recommendations are mine, not the scouts'.

| Rank | Posting | Scout verdict | Actually selecting for | Jordan's evidence (record) | Materials / channel | Strongest case against | Title vs work |
|---|---|---|---|---|---|---|---|
| 1 | Corvid, Research Scientist, Agent Evaluation | Pursue | Someone who can tell whether a grader score means what it claims. No psychometrics or human-subjects person on the team | Relay judge validation against 300 hand-coded messages. Kestrel measure rebuilt and validated against cancellation. Invited Varrow talk on judge trust. JAP 2021 measurement model | PDF résumé via parsing ATS, optional letter read by research team | No RL, training, or computer-use agent work. All AI work since 2024, unreviewed | Mild. Reads as ML role, real gap is measurement validity |
| 2 | Lakeshore, Asst Prof OB and Technology | Pursue | AI-agent experiments and testbeds no faculty member has. Top-journal record. Core MBA OB teaching | PhD OB. *Org Science* 2019, first author. *Mgmt Science* under review. Relay | Seven document types plus three letters via portal. Review begins Nov 1, 2026 | No teaching evaluations, no referees named, no publication since 2021, AI work never involved people | Partial. Teaching is general OB core |
| 3 | Tallgrass, Senior UX Researcher, Measurement | Pursue | First measurement hire to replace engagement counts with validated patient measures | Kestrel rebuild and validation. Parcel experiments and holdout process. JAP 2021 | Résumé plus two PDF case studies via ATS | No patient, usability, clinician, or health-privacy work. Review process not yet built | Yes. IC title, founding-lead scope |
| 4 | Northline, Behavioral Scientist (fractional) | Pursue | Making pricing designs defensible before money is spent. Client-facing credibility | Parcel experiments. Kestrel measurement plans. Pricing rehearsal (simulated only). Working paper | Short email note and résumé to engagement lead. Rate optional | No real-shopper pricing work. No consulting or client-executive record | Minor. Methods review, not research |
| 5 | Meridian, Research Fellow | Ask Jordan | People who have watched government institutions decide. Plain policy writing | PhD. Group-decision papers. AI work on agents, not institutions | Résumé, two-page proposal, general-audience sample via form | No government or policy experience. No general-audience sample exists | Mild. "Institutions" means government |
| 6 | Foundry, Partner | Drop | Fundraiser and founder. 20 to 60 million per organization, boards | Relay matches the thesis. Led three reports | CV plus two 500-word essays via form | No fundraising, board, executive, or budget record | Strong. Science title, fundraising job |
| 7 | Brightwave, Staff DS | Drop | Production ML operations, on-call, Spark and Airflow | Three direct reports, Python, SQL | Résumé via ATS, timed coding test | Fails every hard requirement. No research function | None |

## Decisions

- Pursue: Corvid, Lakeshore, Tallgrass, Northline.
- Drop: Foundry, Brightwave.
- Meridian: drop unless Jordan reports government experience the record lacks.
- Lakeshore is second on fit but first on urgency.

## Notes on scout reports

- All five scouts independently flagged that the record lists AoM 2019 under "Invited" while describing it as competitively selected. Still an open question for Jordan.
- The UX scout calls 2024 to present a gap. The record describes it as independent research, and writing preference 5 says not to raise gaps the documents give no reason to raise. I do not carry that point forward.