Staff ML Engineer, Agent Training & Environments

Posted Yesterday
7 Locations
In-Office or Remote
250K-280K Annually
Mid level
Artificial Intelligence • Information Technology • Machine Learning
Our mission is to build the best products to align with artificial intelligence.
The Role
Build and operate reinforcement-learning environments, agent verifiers, graders, evaluation systems, and fine-tuning pipelines. Run large-scale agent trajectory experiments, develop SFT and RL workflows, and create reliable training and serving infrastructure. The role requires production-grade Python engineering, distributed systems judgment, forensic debugging, compute-cost optimization, and experience improving agentic models through post-training.
Summary Generated by Built In
Shape the Future of AI

At Labelbox, we're building the critical infrastructure that powers breakthrough AI models at leading research labs and enterprises. Since 2018, we've been pioneering data-centric approaches that are fundamental to AI development, and our work becomes even more essential as AI capabilities expand exponentially.

About Labelbox

We're the only company offering three integrated solutions for frontier AI development:

  1. Enterprise Platform & Tools: Advanced annotation tools, workflow automation, and quality control systems that enable teams to produce high-quality training data at scale
  2. Frontier Data Labeling Service: Specialized data labeling through Alignerr, leveraging subject matter experts for next-generation AI models
  3. Expert Marketplace: Connecting AI teams with highly skilled annotators and domain experts for flexible scaling
Why Join Us
  • High-Impact Environment: We operate like an early-stage startup, focusing on impact over process. You'll take on expanded responsibilities quickly, with career growth directly tied to your contributions.
  • Technical Excellence: Work at the cutting edge of AI development, collaborating with industry leaders and shaping the future of artificial intelligence.
  • Innovation at Speed: We celebrate those who take ownership, move fast, and deliver impact. Our environment rewards high agency and rapid execution.
  • Continuous Growth: Every role requires continuous learning and evolution. You'll be surrounded by curious minds solving complex problems at the frontier of AI.
  • Clear Ownership: You'll know exactly what you're responsible for and have the autonomy to execute. We empower people to drive results through clear ownership and metrics.
Role Overview

Labelbox is the RL data factory for advancing frontier agent capabilities. We build the data, environments, and evaluations that frontier labs use to train and judge their agents.

This role sits where training meets infrastructure. You will run the experiments and build the systems that run them: environments agents act in, verifiers that decide whether they succeeded, and the fine-tuning pipelines that turn that signal into a better model. We're looking for someone who does both halves — the engineering throughput of a strong platform engineer, and real depth in post-training agents.

The bar is high: engineers with strong judgment who set technical direction, turn prototypes into reliable systems fast, and are at the frontier of agent-first engineering practice.


What you'll work on
  • RL environments for agentic tasks: task definitions, tool surfaces, state and reset semantics, reward design — and the harness that runs thousands of them in parallel.
  • Verifiers and graders: programmatic checks, LLM judges, rubric pipelines, pass@k scoring. Deciding what "the agent succeeded" means, and making that judgment trustworthy at scale.
  • Fine-tuning pipelines that turn evaluation signals into measurable agent improvements — SFT and RL, from data collection through training to checkpoint evaluation.
  • Eval systems that run millions of agent trajectories to measure model and product quality.
  • Training and serving infrastructure that scales to the throughput frontier labs need: multi-launcher orchestration, long-running job fault tolerance, cost accounting.
What we're looking for

As an engineer

  • A 3+ year track record of shipping systems that customers and other engineers still rely on.
  • Exceptional throughput, without the quality tax. You ship a lot, you review a lot, and the v1 you ship becomes the foundation the rest of the team builds on.
  • Strong system and API design judgment. Hard architecture calls land with you: you make them, defend them under pressure, and update fast when someone else is right.
  • You ship production code with coding agents daily. You know where they break and what it takes to make them reliable, and you use that to move the whole team faster.
  • You build the substrate other people's work runs on — tooling, CI, harnesses, libraries — and you treat that as the job, not a distraction from it.
  • You move fast in ambiguous, startup-pace environments, with influence over authority.
  • Deep proficiency in Python, and comfort across the rest of the stack.

As an RL post-training practitioner

  • You have fine-tuned models for agentic tasks and made them measurably better. SFT plus at least one RL method (GRPO, PPO, DPO, or similar) in production.
  • You have built environments agents operate in, and you know why reward and task design is where most of the difficulty actually lives.
  • You have designed verifiers or graders for open-ended work, and you know how they get gamed.
  • You debug training runs forensically and methodically.
  • You reason about compute-economics. You know what an experiment costs, when a run is not worth finishing, and how to get the same signal for a tenth of the spend.
  • You write up what you learned so it changes what the team does next.

Nice to have
  • Experience with agent harnesses and coding agents as subjects of training and evaluation.
  • Multi-tenancy and isolation for untrusted agent execution: sandboxing, egress control, credential handling.
  • Background in production distributed systems, ML infrastructure, or data systems at scale.
  • Experience working directly with frontier labs or other highly technical customers.

Our Technology Stack

Our engineering team works with a modern tech stack designed for scalability, performance, and developer efficiency:

  • Frontend: React.js with Redux, TypeScript
  • Backend: Node.js, TypeScript, Python, some Java & Kotlin
  • APIs: GraphQL
  • Cloud & Infrastructure: Google Cloud Platform (GCP), Kubernetes
  • Databases: MySQL, Spanner, PostgreSQL
  • Queueing / Streaming: Kafka, PubSub

Labelbox strives to ensure pay parity across the organization and discuss compensation transparently.  The expected annual base salary range for United States-based candidates is below. This range is not inclusive of any potential equity packages or additional benefits. Exact compensation varies based on a variety of factors, including skills and competencies, experience, and geographical location.

Annual base salary range
$250,000—$280,000 USD
Life at Labelbox
  • Environment: Fast-paced and high-intensity, perfect for ambitious individuals who thrive on ownership and quick decision-making
  • Growth: Career advancement opportunities directly tied to your impact
  • Vision: Be part of building the foundation for humanity's most transformative technology
Our Vision

We believe data will remain crucial in achieving artificial general intelligence. As AI models become more sophisticated, the need for high-quality, specialized training data will only grow. Join us in developing new products and services that enable the next generation of AI breakthroughs.

Labelbox is backed by leading investors including SoftBank, Andreessen Horowitz, B Capital, Gradient Ventures, Databricks Ventures, and Kleiner Perkins. Our customers include Fortune 500 enterprises and leading AI labs.

Your Personal Data Privacy: Any personal information you provide Labelbox as a part of your application will be processed in accordance with Labelbox’s Job Applicant Privacy notice.

Any emails from Labelbox team members will originate from a @labelbox.com email address. If you encounter anything that raises suspicions during your interactions, we encourage you to exercise caution and suspend or discontinue communications.

Skills Required

  • At least 3 years of experience shipping reliable production systems
  • Deep proficiency in Python and comfort across the rest of the technology stack
  • Experience shipping production code and using coding agents daily
  • Strong system and API design judgment
  • Experience building tooling, CI systems, harnesses, or libraries used by other engineers
  • Experience fine-tuning models for agentic tasks and measurably improving them
  • Production experience with supervised fine-tuning plus at least one reinforcement-learning method, such as GRPO, PPO, or DPO
  • Experience building environments for agents and designing tasks and rewards
  • Experience designing verifiers or graders for open-ended work
  • Ability to debug training runs methodically and reason about compute economics
  • Experience working in ambiguous, fast-paced startup environments
  • Experience with agent harnesses and coding agents as training and evaluation subjects
  • Experience with multi-tenancy, sandboxing, egress control, or credential handling for untrusted agent execution
  • Background in production distributed systems, ML infrastructure, or large-scale data systems
  • Experience working directly with frontier labs or highly technical customers

Labelbox Compensation & Benefits Highlights

The following summarizes recurring compensation and benefits themes identified from responses generated by popular LLMs to common candidate questions about Labelbox and has not been reviewed or approved by Labelbox.

  • Fair & Transparent Compensation — Pay is considered generally competitive for core roles, and pay transparency is explicitly emphasized. Feedback suggests many full-time employees view their overall pay positively.
  • Healthcare Strength — Comprehensive medical, dental, and vision coverage is highlighted alongside wellness resources and tax-advantaged accounts. These offerings indicate a strong baseline of health benefits.
  • Leave & Time Off Breadth — Unlimited PTO, paid sick days, bereavement leave, and generous parental leave are emphasized. This breadth points to ample time-off support.

Labelbox Insights

Am I A Good Fit?
beta
Get Personalized Job Insights.
Our AI-powered fit analysis compares your resume with a job listing so you know if your skills & experience align.

The Company
HQ: San Francisco, CA
115 Employees
Year Founded: 2017

What We Do

Labelbox is the data factory for generative AI, providing the highest quality training data for frontier and task-specific models. Labelbox’s comprehensive platform combines on-demand labeling services with the industry-leading data labeling platform. The Boost labeling service is powered by the Alignerr community of highly-educated experts, who span all major languages and a diverse range of advanced subjects. They are available on-demand to rapidly generate new data for supervised fine-tuning, RLHF, and more. Labelbox’s software-first approach delivers unmatched control and transparency into the labeling process, leading to the generation of high-quality, consistent data at scale. Customers include Fortune 500 enterprises and leading AI labs, and Labelbox is backed by leading investors including SoftBank, Andreessen Horowitz, B Capital, Gradient Ventures, Databricks Ventures, and Kleiner Perkins.

Why Work With Us

Labelboxers are driven to master their craft and build the best products for AI, which is quickly becoming one of the most significant technologies of our time. Join us in supporting our customers as they create AI breakthroughs.

Gallery

Gallery

Similar Jobs

Samsara Logo Samsara

Consultant

Artificial Intelligence • Cloud • Computer Vision • Hardware • Internet of Things • Software
Easy Apply
Remote or Hybrid
6 Locations
4000 Employees
79K-103K Annually

mTrip Logo mTrip

Full-stack Engineer

Information Technology • Software • Travel
In-Office or Remote
Montreal, QC, Canada
12 Employees

Optum Logo Optum

Program Manager

Artificial Intelligence • Big Data • Healthtech • Information Technology • Machine Learning • Software • Analytics
Remote
Canada
160000 Employees
109K-225K Annually

Samsara Logo Samsara

Customer Success Manager

Artificial Intelligence • Cloud • Computer Vision • Hardware • Internet of Things • Software
Easy Apply
Remote or Hybrid
Canada
4000 Employees
78K-101K Annually

Similar Companies Hiring

Blee Thumbnail
Artificial Intelligence • Marketing Tech • Software
New York, New York
33 Employees
Vega Thumbnail
Artificial Intelligence • Automotive • Insurance • Transportation
US
65 Employees
RapidScale Thumbnail
Cloud • Information Technology • Consulting
Raleigh, NC
600 Employees

Sign up now Access later

Create Free Account

Please log in or sign up to report this job.

Create Free Account