Scale AI has spent a decade being the company that made AI work accessible — the data annotation infrastructure behind almost every frontier model in production, the contractor platform that gave hundreds of thousands of people their first paid AI task, and the RLHF pipeline that Anthropic, OpenAI, Microsoft, and Meta rely on to train the systems that run the world. For the Class of 2026, Scale occupies a position no other company can claim: it is simultaneously the largest entry-level AI employer in the world and one of the most selective full-time engineering hirers in San Francisco.
That duality is the story. No other company in this issue's coverage operates across such a wide band of the hiring market — from a first annotation task at $15 per hour to a new-grad engineering offer in the range of $180,000 to $220,000 in total compensation.
From MIT Dropout to $29B: How Scale Was Built
Scale AI was founded in 2016 in San Francisco by Alexandr Wang, then a 19-year-old MIT dropout who built the company's first data labeling tools himself. By the time Forbes recognized him, Wang was the youngest self-made billionaire on record. The company grew from that foundation — labeling sensor data for autonomous vehicles — into the dominant infrastructure layer for human-supervised AI training across every major model provider.
The May 2024 Series F, backed by Nvidia, Amazon, Meta, Accel, and Index Ventures, valued Scale at approximately $14 billion. Then, in June 2025, Meta acquired a 49 percent stake in a deal that valued Scale at approximately $29 billion. Wang moved to Meta as Chief AI Officer, leading Meta's Superintelligence Labs. Jason Droege, previously Scale's Chief Strategy Officer and a former Uber Eats executive, took the CEO role.
Scale's full-time headcount sits at approximately 1,000 to 1,200 employees across engineering, research, operations, and go-to-market. Its contractor and tasker network — operating under the Outlier and Remotasks platforms — runs into the hundreds of thousands globally. Those two numbers, side by side, define the company's position in the AI labor market.
DoD Contracts, RLHF Pipelines, and Why Every Frontier Lab Is a Customer
The core of Scale's business is data quality at volume: it provides the annotation, evaluation, and RLHF pipelines that turn raw compute and model architecture into working AI systems. Customers include Anthropic, OpenAI, Microsoft, Meta, and the U.S. Department of Defense. The DoD relationship alone is substantial — Scale holds a production agreement with the Pentagon's Chief Digital and AI Office that has a ceiling of $500 million, confirmed by Bloomberg in May 2026, alongside a separate Army R&D contract announced in August 2025.
That government and defense footprint is not incidental to hiring. Scale is building out dedicated teams in Washington D.C. and expanding offices in New York and St. Louis specifically to serve federal customers. The company told reporters in November 2025 that it was looking to hire 200 people across all areas of the business. As of May 2026, more than 400 open roles are listed across Scale's career pages.
Under Droege, Scale's stated mission has shifted from "accelerate the development of AI" to "develop reliable AI systems for the world's most important decisions" — a reframe that signals the company's pivot from pure data infrastructure toward a broader role in enterprise and government AI deployment. The applications business, which includes agentic workflow automation for Fortune 500 clients and defense systems, is projected to double in 2026 on the back of 2025, which Droege publicly described as Scale's strongest financial year ever, with well over $1 billion in new business signed.
In March 2026, Scale launched Scale Labs — an expanded research division built on its earlier Safety, Evaluation, and Alignment Lab. Scale Labs now covers AI model capabilities research, post-training evaluation, enterprise deployment risk, and frontier model safety. For new-grad researchers, Scale Labs represents a direct entry point into applied AI safety and evaluation work.
Two Entry Points, One Company
Scale's approach to entry-level hiring runs on parallel tracks that almost no other company in the market offers.
The first track is the Outlier contractor network. Outlier operates at scale — the company has described its network of annotators as spanning hundreds of thousands of contributors with advanced degrees. Tasks range from generalist content evaluation at approximately $15 per hour to coding RLHF work paying $50 to $65 per hour for computer science specialists to domain expert evaluation in medicine or law at rates well above $100 per hour. For a Class of 2026 computer science graduate who does not land a direct full-time role, a sustained Outlier engagement at the coding tier generates real income — in the range of $50,000 to $75,000 annualized for consistent part-time hours — and, more importantly, a resume line that reads as genuine proximity to production AI systems. Hiring managers at frontier labs have begun treating documented Outlier work, at sufficient volume and quality, as a meaningful signal.
The second track is Scale's full-time new-grad hiring, which operates in two formats.
The Strategic Projects Lead (SPL) program — actively recruiting for its 2026 cohort as of this writing — targets new graduates for operational roles inside Scale's Generative AI business unit. The SPL role is explicitly described as wearing many hats across operations, supporting revenue-generating initiatives, and shaping the company's Gen AI product strategy. Base salary in Scale's hub cities of San Francisco, New York, and Seattle runs $112,000 to $140,000, with equity on top. The program is not purely technical — it draws from STEM, CS, operations research, and data science backgrounds, and it functions as Scale's fast-track path for generalists who want to work at the operational center of an AI infrastructure company.
Scale also carries open Software Engineer - New Grad listings targeting graduates with CS, EECS, computer engineering, or statistics degrees, prior internship experience in product or software engineering, and proficiency in Python, TypeScript, React, and related tooling. Based on available information, Scale's full-time engineering pipeline for new graduates is highly competitive given the company's relatively compact full-time headcount of roughly 1,000 to 1,200 employees — smaller than any of the Magnificent Seven but operating on a client list that spans every frontier lab in the market.
What a First Year Looks Like
A first-year software engineer at Scale AI in San Francisco in 2026 is working inside an organization that serves as the critical quality layer for most of the major models in production. The work is model-proximate by design — Scale engineers build the tooling, pipelines, and evaluation infrastructure that its annotation workforce uses to produce training data.
Compensation for a new-grad SWE at Scale, based on Levels.fyi data, runs approximately $175,000 to $220,000 in total compensation in its first year, combining base salary, equity (stock options granted upon board approval), and benefits including comprehensive health coverage and a learning and development stipend. The company does not have the same equity upside headline as OpenAI or Anthropic — Scale is private and, post-Meta deal, at a valuation that limits early-employee multiples — but the base-plus-options package is competitive with all but the top tier of frontier lab new-grad offers.
For a Scale Labs Research Scientist or Machine Learning Research Engineer, the roles in evaluation, agent robustness, AI controls, and frontier risk assessment are closer in character to an Anthropic research role than to a typical industry engineering position. The research team is small by design; Scale Labs is built to publish, not just to ship.
CEO Jason Droege, speaking on Lenny Rachitsky's podcast in October 2025, described his three-trait hiring screen: curious problem solvers, humble collaborators, and effective leaders. On the question of experience versus adaptability, Droege was direct: "The world's changing, right? So you do need people that are adaptable. So all the experience is not necessarily one-to-one relevant." He carved out one exception — for research roles, he acknowledged, specific domain experience "is critical because the market's moving so fast, you don't have time to train up some people."
That framing maps precisely onto Scale's two-track new-grad architecture. The SPL program hires for adaptability. The engineering and research tracks hire for demonstrated domain depth.
2026 Hiring Trajectory
Scale's 2026 hiring picture reflects three simultaneous pressures. First, the company's government and defense business is expanding faster than its commercial AI infrastructure business, which means Washington D.C.-oriented roles — policy, forward deployed engineering, and government account operations — are among the fastest-growing headcount categories. Second, Scale Labs is building out its research bench with specific focus on evaluation and safety, creating a small but real pipeline for research-oriented new grads who want applied safety work without joining a frontier model lab directly. Third, the SPL program is absorbing new-grad generalist talent at volume — it is the most accessible formal entry point into Scale's full-time workforce for candidates without deep SWE credentials.
The July 2025 workforce reduction — approximately 200 full-time employees and 500 contractors, roughly 14 percent of staff — was a reset driven by what Droege described as over-rapid generative AI capacity expansion and excess organizational complexity. The company has been hiring back since November 2025 into a leaner structure. The $500 million DoD contract ceiling, confirmed days before this issue went to press, is the most direct signal of where new headcount will go.
The $500M Pentagon Signal: Where Scale's Headcount Goes Next
Scale enters the second half of 2026 as a structurally different company than it was 24 months ago. The Meta stake changed its capital position and its principal relationship to the frontier model market. Wang's departure shifted the founding narrative. Droege's "reliable AI" mission reframe is a bet that the next phase of AI revenue — in government, in defense, in enterprise deployment — will be won by the company that can certify and evaluate AI systems for high-stakes decisions, not just annotate training data.
For the Class of 2026, the opportunity set at Scale maps along a clear axis: the contractor network as an on-ramp, the SPL program as a fast-track generalist entry, and the engineering and Scale Labs tracks as the competitive full-time paths for candidates with the depth to clear a selective bar. The company that gave tens of thousands of people their first AI task is now offering a narrower, more deliberate set of doors into full-time AI work — and the $500 million Pentagon contract means those doors are opening wider in Washington than in any other city in its network.
Sources: Scale AI Careers — Software Engineer New Grad — Scale AI Strategic Projects Lead New Grad 2026 — Scale Next Era: Building for 2026 — Scale.com Blog — Scale AI Founder Wang Joins Meta — Scale.com Blog — CNBC: Scale AI Alexandr Wang confirms departure for Meta — Bloomberg: Meta-Backed Scale AI Wins $500 Million Defense Department Deal — Scale AI Jason Droege CEO — Axios — Jason Droege: First Interview Since Taking Over — Lenny's Newsletter — Scale AI CEO: Prioritize Adaptability Not Experience — Global Trade Magazine — Scale Labs Jobs — Scale AI Salaries — Levels.fyi — Alexandr Wang: Building Scale AI — Y Combinator Library — 2026 AI College Jobs — speedyapply GitHub