Surge AI

What is Surge AI?

Surge AI supplies the human feedback and data annotation that AI labs use to train and evaluate large language models, a practice known as RLHF (reinforcement learning from human feedback). Its contributors rate, correct, and refine model outputs, build RL environments, and label data so models learn nuance, tone, sarcasm, correctness, and domain knowledge that raw internet text alone doesn't teach them. Clients have included OpenAI, Anthropic, Meta, and Microsoft and Google, and the company has said its work contributed to training models such as Claude and Gemini. Its differentiator versus many data-labeling competitors is leaning on highly qualified "expert" contributors, people with advanced degrees or deep professional or native-language expertise, rather than low-skill microtasks, and paying accordingly. Surge AI (legal entity Surge Labs Inc.) was founded in 2020 by Edwin Chen, CEO, a former data scientist at Google, Twitter, and Facebook and an MIT dropout. It is headquartered in San Francisco and was bootstrapped for years without outside investment, reporting roughly $1.4 billion in 2025 annual recurring revenue. In mid-2025 it was reportedly in talks for its first external capital raise, up to $1 billion, at a valuation reported anywhere from around $15 billion to $25 billion depending on the source; that raise had not been confirmed as closed as of this writing. The company employs roughly 130 full-time staff, plus a much larger global network, reported at around 50,000, of contract annotators and expert contributors. Separately, Surge AI faces a worker-misclassification class-action lawsuit in San Francisco Superior Court, alleging it misclassified contract annotators as independent contractors, required unpaid training time, and denied California labor-law protections like minimum wage, overtime, and meal/rest breaks. This is a pending, unresolved allegation, not a proven finding, but relevant context for contract annotation work.

Mission & values

Provide the high-quality, expert human feedback that helps AI models understand nuance and become genuinely useful and safe.

Qualifications

Surge AI has two different kinds of work, and it matters which one you're applying for. The small full-time corporate team hires software engineering, machine learning and infrastructure, product, operations, and go-to-market roles to run the platform and client relationships; these are salaried positions based around its San Francisco headquarters. The much larger group is its global network of contract data annotators and "expert" contributors: people who rate, correct, and label AI outputs, build RL environments, and complete specialized annotation tasks. Unlike many competitors, Surge AI leans heavily toward contributors with advanced degrees, professional credentials, or deep subject-matter or native-language expertise, medical, legal, coding, and specific-language experts are in particular demand, and it is not really designed for beginners. Contributor work is remote, task or project-based, and paid per minute or per task rather than a salary; reported baseline rates for approved contributors run roughly 30 to 40 cents per working minute (about $18 to $24/hour), while highly specialized expert work has reportedly paid several hundred dollars per hour in some cases. Note the pending worker-misclassification lawsuit described above when weighing contract annotation work here.

Leadership

Hiring process

1. 1

   Apply at surgehq.ai

   Corporate roles: browse engineering, ML, product, and operations openings. Contributor/expert roles: apply to the contributor program, often specifying your domain expertise, credentials, or language fluency.

2. 2

   Screening and qualification

   Contributors typically complete a qualification process demonstrating relevant expertise (subject-matter, professional credential, or native-language fluency); the bar is high and not aimed at beginners.

   About 7 days

3. 3

   Sample or paid trial tasks

   Approved contributors are given initial tasks or projects to establish quality and reliability before being offered more, or higher-paying, work.

   About 7 days

4. 4

   Ongoing task/project work

   Contributors pick up available annotation, RLHF, or expert-review tasks; volume and pay depend on your qualification tier and current client demand.

5. 5

   Payment

   Paid per task or per working minute for contributor work; corporate roles are on standard salaried payroll.

   About 7 days

Funding

Awards & recognition

• Reached ~$1.4B in 2025 annual recurring revenue, fully bootstrapped · 2025

  Surge AI

• Provided RLHF/data used in training models including Claude and Gemini · 2026

  Surge AI

Company information

• Industry: AI Training Data / RLHF
• Location: Remote (contractor tasks, global); HQ San Francisco, CA
• Founded: 2020
• Employees: 130
• Website: surgehq.ai ↗
• LinkedIn: Company page ↗

Frequently asked questions