Distinguished Researcher.

The lab had reached the point where its senior research bench needed a clearly defined IC apex above the staff and principal levels. The Distinguished Researcher seat was designed to anchor a small group of senior ICs, set the technical direction of one of the lab's research lines, and represent the lab externally on substance — at conferences, in policy fora and in the lab-to-lab back-channel that quietly shapes the sector. The candidate had to be a researcher of the first rank, not an organisational figure standing in front of one.

The non-negotiables were narrow and exacting. Published work cited inside frontier training teams, a research line with technical coherence rather than breadth, and a personal posture suited to an IC seat rather than a managerial one. The comp band was structured at the upper edge of senior research compensation, with equity that reflected the strategic weight of the hire. Geography was preferred near the lab's research hub, with willingness to consider candidates relocating. The CEO was explicit that the lab did not want a senior researcher who had drifted into management — they wanted a researcher.

The senior IC bench in foundation research is one of the most observed and most slowly moving segments in AI. The candidates who hold the apex are few, well-known to one another, and rarely in the open market. Reference networks are tight; movement between organisations is read inside the sector within days, and the implications of a researcher's move are debated as a market signal in their own right. Most senior researchers at the level the lab was hiring for were either tenured inside their current organisation by equity, by ongoing research lines, or by the gravitational pull of their existing research group.

Our read was that the shortlist would have to be constructed candidate by candidate, with each conversation taking longer than a conventional executive search would tolerate. Search firms without practitioner depth in this segment cannot read whether a candidate's contribution to a frontier training run was load-bearing or peripheral — and the lab's senior researchers would read that distinction instantly. We assessed each candidate on the specific technical contributions in their published work, on the citation patterns inside the lab itself, and on the back-channel signal from researchers who had collaborated with them directly. The conversations were slow by design.

The shortlist composition was deliberately narrow. We weighted depth on a coherent research line above breadth across many, and we screened heavily for the IC temperament — researchers who wanted to do the work themselves, not to direct others doing it. Each candidate was assessed on their own published artefact rather than on the seniority of their current title.

• Senior researcher with a coherent frontier research line.
• Staff scientist whose recent work was cited inside the lab.
• Researcher returning to an IC seat after a brief management chapter.
• Academic-adjacent researcher with frontier-lab collaboration history.

The hire came from the narrowest band of the shortlist — a senior researcher whose recent published work was sitting on the desks of the lab's existing research leads, and whose technical posture was unambiguously IC rather than managerial. What made them right was the integration of contribution and temperament: the lab's senior researchers had been engaging with their work for months, and the candidate themselves arrived at the brief with a clear, specific view of which research line they wanted to anchor. The Head of Research described the final interview as the first time the seat had felt necessary rather than aspirational.

The close was the slowest of any AI engagement we ran this year, by design. The candidate held parallel conversations with two adjacent labs and required a research-line commitment from the Head of Research that was specific enough to be operationally meaningful. We worked through the structuring of the seat across three weeks, with the CEO joining the final conversation on long-horizon research commitments. Brief to offer ran sixteen weeks, with the offer accepted in the second of two reference cycles.

“The engagement reinforced a discipline we now hold for the apex of the senior IC bench: this is not an executive search dressed as a research one, and the cadence should not borrow from that template. The conversations are slower because the assessment is on the specific research artefact, the back-channel signal, and the candidate's own technical convictions rather than on competencies. The hire was the candidate whose published work, internal lab reception and personal posture were already aligned before the search began — the role of the search was to recognise that alignment and to construct a seat that the candidate could move into without compromise.”

— Craig Oliver