Agentic Tasks & RL Environments

If you can take a messy real task, pin down exactly what "done correctly" means, and turn that into a check a machine can run thousands of times, you are qualified for the best-paid and fastest-growing corner of the entire AI-training economy. Building the interactive tasks and reinforcement-learning (RL) environments that agentic models train against sits near the top of the pay ladder (its open roles on this site typically pay Unavailable to Unavailable an hour today, with a median of Unavailable) for one blunt reason: it is not labeling, it is software and task engineering, and the frontier labs are spending on it at a scale that dwarfs the rest of the field. Anthropic alone was reported to be discussing $1 billion or more over the following year on environments (TechCrunch, 21 Sep 2025), and an industry tally posted on 12 Jul 2026 counts 50-plus vendors selling data and RL environments to labs at roughly $8.5 billion in combined annual revenue. This guide covers what the work actually is, who buys it, what it really pays per task and per hour, what you genuinely need to do it, and the open on-ramp that lets anyone in the world build a portfolio the labs can see.

The honest headline up front is that this is a two-sided market. At the top, the startup Mechanize reportedly offered $500,000 salaries to engineers who build RL environments (TechCrunch, Sep 2025), and single authored tasks change hands for $200 to $2,000 each (Epoch AI, 12 Jan 2026), with a full product replica reaching six figures. At the bottom, the median AI training role on this site advertises Unavailable an hour, and data labeling roles typically start around Unavailable. The gap between those two worlds is precisely what this guide is about: the skill of designing tasks and writing verifiers, not the willingness to click.

What "Agentic Tasks" and "RL Environments" Actually Are

Modern models are increasingly trained to act, not just answer. They run code, click through applications, search files, and complete multi-step tasks. Teaching that behavior needs a fundamentally different human input than rating two chatbot answers. According to Epoch AI's authoritative FAQ on the subject, an RL environment has three parts: a set of actions the model can take (run code, call a tool, edit a file), the surrounding state that determines what those actions do (a file system, an application, a simulated app), and a grader that scores each attempt. Scored attempts update the model's weights so it repeats what worked. One founder described building one to TechCrunch as "creating a very boring video game": a small, rule-bound world with a win condition.

RLVR: the shift from preference to proof

The paradigm underneath this is RLVR, reinforcement learning with verifiable rewards. Where classic RLHF trains a learned reward model from human preferences (a person picks the better of two answers, and a second model learns to imitate that taste), RLVR gives the model a reward only when its output passes an automated, deterministic check: the unit tests pass, the math answer matches, the database ends in the correct state. As Appen explains, the reward is computed, not predicted. The advantage is structural: there is no learned reward model to hack, because a deterministic verifier cannot be flattered. RLVR started where answers are cleanly checkable (math, code, logic) and in 2026 is being pushed into agent workflows (tool use, computer use, enterprise tasks) where, as Daily Dose of DS notes, writing a trustworthy verifier is much harder and the human judgment is worth much more.

The vocabulary you need

A handful of terms recur across every listing, and knowing them is the difference between reading a job post and understanding it:

• Trajectory (or rollout). One full attempt by the agent: its complete sequence of actions and observations. Labs generally want a minimum pass rate of around 2 to 3 percent, or at least one success in 64 or 128 rollouts (Epoch AI, 12 Jan 2026). A task the model never solves may be broken or infeasible; a task it always solves teaches nothing.
• Grader (or verifier, or rubric). The code or expert-written rule that scores the attempt. Robustness against reward hacking was cited as the single most important quality criterion by the environment founders Epoch surveyed, because a model will find and exploit any loophole in a sloppy check.
• Gym. A colloquial name for an interactive environment. SemiAnalysis reported on 6 Jan 2026 that OpenAI has purchased hundreds of website replicas ("UI gyms") to train ChatGPT Agent to operate software.
• verifiers / prime-rl. Prime Intellect's open-source library and trainer for building environments with built-in reward functions, and the most accessible way for an individual to actually author one (more on that below).

Where the demand is concentrated

Epoch groups the demand into three domains: math and coding (the original focus, with math emphasis now declining), enterprise workflows (the major growth area: expense reports, spreadsheets, CRM navigation, multi-tab office tasks), and longer-horizon multi-step tasks. Coding is widely treated as the "holy grail" vertical, and the messy enterprise-workflow environments are the fastest-growing, which matters for where a newcomer should aim.

What You Actually Do, Day to Day

Not all of this work requires you to engineer a sandbox from scratch. The category spans a ladder from an accessible on-ramp to deep specialization, and most contracts mix several of these tasks. Per Epoch and the Wing Venture Capital market analysis, contributors design tasks, write the reference solution, specify the reward signal, containerize the environment, and in some cases grade the model's rollouts and QA the whole thing against reward hacking.

The through-line is the same across the ladder: your product is a decision about what correct means and a mechanism that enforces it. A trajectory annotator decides whether a given step was a mistake; an environment builder decides what the whole task even rewards. Both are judgment about correctness, encoded so a machine can act on it repeatedly. That is why the work resists automation and why it pays what it does.

To make it concrete, consider a coding environment built end to end. You take a real bug in an open-source repository, write the task ("make the failing test pass without breaking the others"), and set the grader to the existing test suite itself, so pass or fail is decided by code rather than opinion. You containerize the repo so an agent can check out the code, edit files, and run the tests in a sandbox, then you tune the difficulty until the model succeeds only rarely, near the minimum pass rate Epoch says labs want. Finally you attack your own grader: can the model make the test pass by deleting it, hard-coding the expected output, or weakening an assertion? Closing those holes is the actual craft, and it is why a well-built environment is worth far more than a raw bug report.

Who Buys It, and What Shape the Market Is In

The buyers are the frontier labs (Anthropic, OpenAI, Google, Meta, xAI) and a growing set of product companies. Some, such as Cursor, build environments in-house; others, per Epoch, partner with labs on replicas of their own products (Salesforce, Slack, Shopify, and Stripe among them), so a model can learn to operate the real software.

The vendor market has ballooned. The tally, posted by investor Deedy Das on 12 Jul 2026, counts more than 50 companies selling data and RL environments to labs, roughly $8.5 billion in combined annual revenue and about $100 billion in combined valuation, with more than 75 percent concentrated in four players: Scale, Surge, Mercor, and Handshake. For context, Wing VC (January 2026) puts the older, traditional labeling market at about $5 billion growing more than 50 percent a year, and predicts the roughly 20 seed-to-Series-A RL-environment startups will consolidate to 3 to 5 leaders by 2030. Treat the $8.5 billion and $100 billion figures as an analyst estimate rather than an audited number, but the direction is not in dispute.

The vendors, by verified status (mid-2026)

• Mercor pivoted hard from talent marketplace into RL environments and domain-expert data. It raised a $350 million Series C at a $10 billion valuation in October 2025, and Sacra estimates it reached roughly $2 billion in annualized gross revenue by June 2026, and Forbes reported on 9 Jul 2026 that it was in talks to raise at a $20 billion valuation. In February 2026 it acquired Sepal AI, a platform building training data, eval benchmarks, and RL environments for frontier LLMs, to "train models against the most complex tasks humans perform."
• Surge AI is bootstrapped and large: reportedly about $1.2 billion in 2024 revenue, with Bloomberg reporting on 30 Jul 2025 that it was in talks to raise about $1 billion at a valuation of at least $25 billion (talks; no close has been reported). It built a dedicated internal RL-environments team and ships EnterpriseBench, a simulated startup world where frontier models solve only about 30 percent of tasks (Surge, Feb 2026), after even GPT-5 and Claude Sonnet 4.5 failed over 40 percent of agentic tasks in an earlier Surge environment (Nov 2025), numbers that quantify the entire opportunity.
• Scale AI / Outlier is the cautionary tale. After Meta took a roughly $14.3 billion stake in June 2025, rival labs cut contracts and Scale laid off about 14 percent of staff (TechCrunch, 16 Jul 2025). Its then-CEO Jason Droege told Forbes in May 2026 that Scale expects to top $1 billion in revenue in 2026, after just under $1 billion in 2025, half the $2 billion it had projected for 2025 before Meta's deal (Bloomberg, April 2025). That shortfall is the clearest single illustration of the shift away from static labeling.
• Prime Intellect is the open on-ramp. It launched its Environments Hub in August 2025, a community platform its homepage says holds more than 2,500 community environments (read 29 Oct 2026), and it ships the open-source prime-rl trainer and verifiers library. Its paid Environments Bounty Program (an open tier of $100 to $500 per environment and an application-only tier of $1,000 to $5,000-plus) now says it is closed to new applications and claims (read 29 Oct 2026); publishing to the Hub stays open worldwide.
• Turing repositioned around "AGI advancement," raising a $111 million Series E at a $2.2 billion valuation in March 2025, and built a 10,000-plus Python-task dataset (Feb 2026) for reinforcement learning with execution feedback.
• Handshake AI pivoted from a college-jobs platform to AI data and reached about $1.1 billion in annualized gross revenue by April 2026, about $950 million of it from AI data work, by Sacra's estimate (22 Apr 2026), one of the big four, acquiring the quality-tooling firm Cleanlab in January 2026 to move into evaluation and QA.
• Micro1 went from about $7 million in annual revenue at the start of 2025 to more than $500 million by September 2026, on the strength of its AI recruiter, when Forbes reported it had raised over $100 million at a $4 billion valuation (22 Sep 2026); Sacra estimates it reached $1 billion annualized in October 2026.
• Mechanize, founded in April 2025 by ex-Epoch researchers to build RL environments for full economic automation, is working with Anthropic and offered $500,000 salaries to environment engineers (TechCrunch, Sep 2025); Tracxn reports about $9.1 million raised over two rounds (read 29 Oct 2026).
• Invisible Technologies raised about $100 million in September 2025 (Inc., 18 Sep 2025), earmarking funds for RL gyms and "mirror worlds," digital twins of real operations.
• Alignerr (over Labelbox) is the contributor-facing marketplace; its rate card on this site, checked 6 Nov 2026, puts its roles at $10–$200/hr (German voice actors to $280/hr).

Newer pure-play entrants that Wing names include Bespoke and Applied Compute, and directories such as rl-list.com track the growing field.

What It Pays

Pay in this category splits into two very different numbers: what the lab pays the vendor per task, and what the vendor pays you per hour. Understanding both keeps you from either underselling your work or believing a headline that is really an enterprise contract price.

Per-task pricing (what labs pay vendors)

Per Epoch AI (12 Jan 2026), individual expert-authored tasks run $200 to $2,000 each (a rare, very hard one can hit $20,000). A website replica costs on the order of $20,000 (Epoch citing SemiAnalysis), and one founder Epoch interviewed put a complex product replica such as a Slack clone at around $300,000. Exclusive deals cost roughly 4 to 5 times a non-exclusive one, and full lab contracts are often six to seven figures per quarter. These are vendor-level prices, not your paycheck, but they explain why the hourly rates below are high: the underlying work is valuable.

Hourly pay to the worker

Sources: the entry, mid-level and expert rows are from Mercor's own rate guide (updated 23 Jun 2026), which also puts the national average AI-trainer rate at $31/hr as of early 2026; the rows "on this site" are read live from the open roles when you load this page. Mercor CEO Brendan Foody said on 6 May 2026 that the platform's average pay rate had "surpassed $100/hour", and its homepage showed an average contracted rate of $90/hr on 8 Nov 2026. For domain experts, AI Gig Jobs puts PhD AI-training work at $80 to $200/hr and direct lab contracts at $150 to $300/hr (Mar 2026), and PhDs and grad students are described as the highest-paid tier by a wide margin. Separately, bounty programs pay per accepted environment: Prime Intellect's paid $100 to $500 on its open tier and $1,000 to $5,000-plus by application before it closed to new claims (read 29 Oct 2026), which behaves more like piecework and portfolio-building than an hourly job.

The honest caveat: the $500,000 salaries and the $300/hr direct lab contracts are the top of a wide distribution, not the median, which for agentic roles on this site is Unavailable an hour today. Plan around the expert band and treat the headline as the ceiling.

What a realistic engagement looks like: part-time or full-time contract work that arrives in project waves rather than on a fixed schedule. A credentialed engineer holding steady expert-tier work earns the rate times the hours the project actually gives, and like all contract work the volume is lumpy and the effective rate sits below the headline once you count ramp-up and unpaid calibration. Bounty work behaves differently again: it pays per accepted environment as a one-off, which makes it a portfolio investment as much as an income stream.

What You Need to Do This Work

This is where agentic-eval work separates cleanly from generic labeling. The premium roles require real software engineering: writing Python, building containerized environments with Docker, and authoring unit-test-style graders that resist reward hacking. Because reward functions are custom code, you often cannot tell whether the function measures what you think until after training, so debugging and verification skill matter as much as the initial build.

Alongside the engineers, labs pull in domain experts to define domain-specific tasks, write reference solutions, and specify reward signals: finance professionals for financial tasks, doctors for medicine, lawyers for legal reasoning. Machine-learning literacy helps but is not always required for domain-expert authoring, because the labs supply the training machinery; what they cannot supply is your judgment about what a correct outcome is. The contrast with RLHF preference-rating is sharp: rating needs consistent judgment against a rubric, while environment authoring needs the ability to build and test executable software with a correct automated grader, or the equivalent expert reasoning trace in a licensed domain.

In practice the single most valuable engineering skill here is verifier debugging. Because a reward function is just code, a subtle mistake can quietly reward the wrong behavior for an entire training run before anyone notices, which is exactly the failure mode that made robustness against reward hacking the top-cited quality criterion. Engineers who can write a grader, adversarially test it against a model that is actively trying to cheat, and then confirm after training that it measured what they intended are precisely what the labs are short of.

How to Break In (Including the Open On-Ramp)

Unlike most premium work, this category has a genuinely open door that requires no application at all.

1. Build a public environment on Prime Intellect. The Environments Hub lets you build and publish an environment using the open-source verifiers library and create a public artifact a lab can inspect; its paid bounty program says it is closed to new claims (read 29 Oct 2026), so treat the Hub as a portfolio route. Prime Intellect crowdsourced more than 400 frontier environments in about two months this way (27 Oct 2025). HUD.ai is another route to publish environments to labs.
2. Apply to the expert marketplaces. Mercor (an AI-vetted interview), Surge AI, Handshake AI, Turing, Micro1, and Alignerr all recruit for this work; browse the live openings above for what is posted right now.
3. Expect a real screen. A technical or coding assessment for engineering roles, credential verification for domain experts, then trial tasks, and part-time or full-time project contracts.
4. Lead with the right portfolio signal. A published RL environment on GitHub or the Environments Hub, demonstrated Python plus containerization plus a clean grader that resists reward hacking, and domain credentials for the specialist tracks.

The strategic point: the on-ramp and the credential are two different doors. A strong engineer without a PhD gets in through built artifacts; a physician or lawyer gets in through the credential and a reasoning trace. Aim at the door that fits what you already have.

Common Misconceptions

• "It is just the next word for labeling." No. Commodity labeling is being automated and its rates are falling, while designing tasks, graders, and reasoning traces is priced at $200 to $2,000 per task (Epoch AI, Jan 2026) and $75 to $200/hr for expert work (Mercor, Jun 2026). It is the opposite end of the same market.
• "You need a PhD." Only for the credentialed tracks. Law, medicine and finance roles on this site typically pay Unavailable to Unavailable an hour, but strong software engineers without a PhD are in high demand to build environments, and the Prime Intellect Environments Hub is open to anyone worldwide.
• "The work is drying up." Not this category. Static labeling is not where the growth is ( Scale expects 2026 revenue just over $1 billion, roughly flat on 2025 and half the $2 billion it had projected for 2025 before Meta's deal) while RL environments are the growth area, with Anthropic discussing $1 billion-plus over a year (Sep 2025).
• "AI will just build its own environments." That is Mechanize's long-term thesis, but 2026 demand for human-built environments is rising, not falling, because the bottleneck is quality at scale, not availability of tooling.

The 2026 Outlook

Frontier labs are spending heavily because RL environments are the binding constraint on the next capability jump. Models are being trained to act, and that requires interactive, verifiable environments rather than static text. Epoch reports (12 Jan 2026) that RL is the dominant use of these environments (one founder put requests at 10 to 20 times those for benchmarking) and that enterprise-workflow environments are a major growth area. The stated "number one bottleneck" is maintaining quality while scaling: contractor motivation, verification, and anti-reward-hacking QA, all of which keep human labor central even as tooling improves. Prime Intellect frames closed, expensive environments as the reason open models fall behind, which is the motivation behind its open Hub.

The realistic forecast for a worker: expect continued consolidation ( Wing sees 3 to 5 winners by 2030, January 2026), continued lab in-sourcing, and rising demand for people who can build and verify environments, especially in coding and messy enterprise workflows. The floor of generic labeling will keep eroding; the ceiling for people who can specify correctness will keep rising.

Bottom Line

Agentic tasks and RL environments are among the highest-paid, fastest-growing, and least understood corners of AI-training work in 2026. If you can engineer a sandbox and write a grader that a model cannot cheat, or if you hold real domain expertise and can author the reasoning a model should learn from, this is some of the most durable and best-compensated work in the field: its open roles on this site typically pay Unavailable to Unavailable an hour today, and labs pay vendors $200 to $2,000 for each authored task (Epoch AI, 12 Jan 2026). The open door is real: build one environment on Prime Intellect's Environments Hub, publish it, and you have an artifact the labs can see. If your background is generalist today, treat this as the direction to grow toward: start with rating and trajectory annotation, build genuine depth in one domain, and move up into graders and environments as your track record grows.