Scale AI’s Nov 7 FDE board: 14 Forward Deployed titles, eval harnesses written into the enterprise role
A 7 November AI Native read of Scale AI’s Greenhouse board found 14 titles with “Forward Deployed” plus two Deployment Strategists; enterprise Frontier Agents Engineer bands list base pay from $180K–$315K by level, and the IC job names evaluation harnesses, guardrails and observability as core work.
On 7 November 2026, AI Native checked Scale AI’s Greenhouse careers page (168 open roles). Fourteen titles contained “Forward Deployed,” and two more were Deployment Strategist roles. Nine of the fourteen were public-sector facing: one US Public Sector Forward Deployed Software Engineer, seven Global Public Sector full-stack Forward Deployed roles (Kuwait City, Riyadh, London, Doha), and one Public Sector Forward Deployed Product Manager. There was no Japan location on the board that day.
Enterprise work is posted as Frontier Agents Engineer (IC / Senior / Staff). The IC write-up puts evaluation and production quality in the job itself—offline benchmarks, golden datasets, regression suites, LLM-as-a-Judge harnesses, tracing, monitoring, guardrails and grounding—alongside integration into customer clouds, warehouses and APIs. Experience floors are 4+ years (IC), 5+ (Senior) and 8+ (Staff). Published US base ranges for those three Forward Deployed enterprise titles are $180,000–$225,000, $216,000–$270,000 and $252,000–$315,000.
Public-sector postings add clearance and travel detail: US Public Sector asks for active TS/SCI or eligibility, with travel often cited at 25% and up to 50%; several GPS Senior roles in Riyadh and Kuwait City say up to two weeks a month in client offices, and Arabic appears as a nice-to-have on six of seven GPS engineer ads (not required). Scale’s Deployment Strategist (e.g. DHS/DOJ) is written as a quota-carrying account leader—not the Palantir-style discovery DS—so the same title means different work by employer.
Compared with same-day OpenAI and Palantir FDE ads, Scale’s enterprise posting is the most explicit about measuring agent quality after go-live. That is the signal for FDEs and for companies writing their own FDE job descriptions: delivery ownership now includes the eval and observability loop, not only the first production cutover.
Source: www.ai-native.jp