Scale AI

What Scale AI does

Scale AI provides the data-labeling and RLHF infrastructure that trains foundation models — historically the backbone behind GPT-4 and Gemini data pipelines, and several US military AI systems. Its business has pivoted meaningfully toward government and marketplace offerings: Scale Donovan targets the public sector, and Scale Evaluation, released April 2025, benchmarks LLMs. The company's defining recent event is financial and structural: on June 12, 2025, Meta acquired a 49% non-voting equity stake for $14.3B, taking founder-CEO Alexandr Wang with it. That deal set Scale's post-money valuation at roughly $29B and lifted cumulative funding into the $14.8B-$15.9B range. For candidates, the key nuance is that Scale is really two companies fused together: a large operational data-labeling business (heavily contractor-driven) and a smaller, senior-heavy software/ML engineering org. The Meta investment reshaped both the cap table and the strategic direction toward enterprise and government, and understanding which side of the house a role sits on matters enormously.

What it is like to work at Scale AI

Culture reads very differently depending on whether you are engineering staff or field/labeling staff. Reviews describe a "fast-paced work environment at the forefront of AI innovation," but the negatives are blunt: "very unorganized from the top down," "highly clique-ish," "pay is low," and "few opportunities for advancement" show up repeatedly, especially in the St. Louis labeling-hub reviews. The aggregate scores capture the split: Glassdoor sits at 3.6/5 across 411 reviews (career opportunities 3.8, culture/values 3.1, work-life balance 2.9), while Indeed is a harsher 2.5/5 across 103 reviews with 64% recommending. Blind pros cite "interesting problems, smart people, lots of new problems"; the dominant Blind con is "uncertainty from leadership." That leadership-uncertainty theme is directly tied to the post-Meta reorganization. Bottom line — engineering roles can be genuinely stimulating, but organizational churn and a wide experience gap between role types are real and well-documented.

What Scale AI pays

Software-engineering comp is competitive, but it skews heavily toward senior levels. Levels.fyi shows an SWE range from $234K at L3 to $577K at L6, with a median roughly around $351K across levels. Because cash flows disproportionately to senior research and platform engineers, new-grad comp reportedly posts below frontier LLM labs like xAI — a candidate expecting frontier-lab new-grad numbers may be surprised. Equity is anchored to the $29B post-Meta valuation, but that number carries real complexity: Meta's 49% non-voting stake and the resulting questions about customer independence and future liquidity make the equity harder to underwrite than a clean pre-IPO story. The practical advice is to separate the labeling-operations pay conversation (frequently criticized as low in reviews) from the SWE conversation entirely, and to press hard on level, equity structure, and how the Meta relationship affects any future secondary or exit path before anchoring on a headline total-comp figure.

How hiring works at Scale AI

The standard loop runs a recruiter screen, a technical phone screen, a hiring-manager round, and a final virtual day with three to four loops. For ML-oriented roles, candidates report a more specialized shape: two coding rounds — one focused on data parsing and statistics, one on ML coding — plus ML deep dives on top of the standard behavioral and system-design components. That structure reflects Scale's identity as a data-and-ML-infrastructure company rather than a pure product shop, so preparation should emphasize applied data manipulation, statistics, and practical ML implementation as much as conventional algorithms. The process is fairly standard in length and rigor for an infrastructure company and is not reported to be as brutal as Glean's ~3%-pass-rate gauntlet. Candidates who can comfortably move between data-wrangling, statistical reasoning, and clean coding under time pressure — rather than only competitive-programming puzzles — will be best positioned across both the SWE and ML-engineer tracks.

Growth & trajectory

Scale's headcount trajectory is a sharp reset followed by selective rebuilding. Its pre-layoff full-time workforce was around 1,400, and on July 16, 2025 it cut 14% — roughly 200 full-time staff — largely in data labeling, and also cut ties with about 500 global contractors. That came just weeks after Meta's June 2025 investment. Since then, the pattern has been a hard cut in the commoditizing labeling business followed by targeted enterprise and government re-hiring around Donovan and Scale Evaluation. So the net headcount story is not simple growth — it is a strategic re-composition, shrinking the contractor-heavy operations while building out higher-margin public-sector and evaluation product lines. For candidates, the trajectory question is really about which part of Scale you are joining: the parts being wound down, or the enterprise/government parts the company is actively investing in post-Meta. Those have very different growth outlooks.

Risks to know

The dominant risks are structural and legal. Founder-CEO Alexandr Wang left for Meta on June 12, 2025 as part of the $14.3B deal, and Meta's 49% non-voting stake raises pointed questions about data-pipeline independence — a real concern for customers who compete with Meta and a source of the "leadership uncertainty" theme in reviews. The July 2025 layoffs hit data labeling hardest, and the broader business faces commoditization pressure in labeling. On the legal side, the Schuster PTSD class action, filed August 11, 2025, alleges that data-labelers were exposed to traumatic content — a reputational and liability overhang tied to the core operational model. Taken together, the risk stack is a founder-CEO departure, a controlling-adjacent minority investor with competing interests, workforce cuts, and active litigation. None is necessarily fatal, but candidates optimizing for a stable, founder-led narrative should recognize that Scale's story has fundamentally changed since mid-2025.

Who thrives at Scale AI (and who should not)

Thrive: candidates who are comfortable with contractor-heavy operations, US government and DoD pipelines, and continuously reorganizing around enterprise swings like Donovan and Scale Evaluation. People who like ambiguity, want to help build new product lines in the wake of a strategic pivot, and can operate through leadership uncertainty will find real room to make an impact — especially on the higher-margin public-sector and evaluation teams the company is actively investing in. Avoid: candidates who want a stable founder-CEO narrative (Wang is gone to Meta), a low-risk and predictable customer progression, or compensation parity with frontier LLM labs — cash skews toward senior engineers, new-grad comp trails, and the post-Meta direction is still settling. It is also a poor fit for anyone uneasy about the active PTSD litigation, the labeling-business commoditization, or the independence questions raised by Meta's 49% stake. Match yourself carefully to the specific team, not the Scale brand.

Roles Scale AI is hiring for

The roles Scale AI is most actively hiring right now in our index, with a live count and the salary guide for each:

The full board of open roles — with comp and location on every posting — is at the top of this page.

The signals behind this page

The hiring picture here is read from 128 live Scale AI postings in our index (refreshed weekly); 66% are remote-friendly, and in a recent sample 64 disclose a pay-transparency band. The culture, growth, and interview detail above is researched and cited; the open-roles board is live from our jobs index.

Sources

• Scale AI $29B valuation & pipeline profile (tsginvest, 2026)
• Scale AI founder Wang exits for Meta, $14B deal (CNBC, Jun 12, 2025)
• Scale AI lays off 14% of staff, mostly labeling (TechCrunch, Jul 16, 2025)
• Scale AI to cut 14% following Meta investment (Bloomberg, Jul 16, 2025)
• Scale AI Software Engineer Salary $234K-$577K+ (Levels.fyi)
• Get a Job at Scale AI: interview process (Exponent)
• Lessons from Schuster v. Scale AI (Workforce Bulletin, Aug 11, 2025)