Above the typical $281k a year for staff-level Machine Learning Engineer jobs in CaliforniaHow we compare
How we compare. Typical is the median yearly pay of 158 staff-level Machine Learning Engineer jobs in California that list pay, from the employers' own postings. About means within 5%.
Run experiments on how environments shape agent behavior, including reward hacking, task cheating, unsafe tool use, and deception, then improve the environments.
Test graders for ways agents can score without doing the task, fix vulnerabilities, and maintain attacks as recurring tests.
Build trajectory monitors, red-team harnesses, and safety benchmarks for real agent runs.
Secure environments through network isolation and controls against escapes, credential misuse, and agent-caused damage.
Summarised from the posting by AI and can be wrong. The full description is the source.
Requirements
Experience
Experience putting research into production systems, especially post-training, distillation, or evaluation
Published work on control, dangerous capability evaluations, oversight, or interpretability (preferred)
RLHF or RLAIF experience and a view on how training choices affect behavior (preferred)
Built or operated a public benchmark or agent task suite (preferred)
Experience running many agents at once using sandboxes, containers, and logging (preferred)
Advised on AI safety or governance policy (preferred)
Skills
Build your own test pipelines and learn unfamiliar codebases quickly
In-depth knowledge of LLM safety, including attacks, red teaming, monitoring and control evaluations, and reward hacking
Turn concerns about model behavior into experiments and explain what the results do and do not prove
Write clearly enough that research results change what others do
Make decisions quickly with incomplete evidence and defend them with data
Carefully review transcripts and identify details that could invalidate results
Your match
Sign in and add your resume to see which of this job's requirements it shows.
Saved to your account, encrypted. The AI that reads it doesn't keep it or train on it.
Benefits
Health insurance
Health benefits
Dental & vision
Dental and vision benefits
Paid time off
Flexible PTO
Equity
Generous early-stage grant
Bonus
Sizable performance bonus tied to project and company milestones
Full job description
From Abundant’s careers site
ABOUT ABUNDANT
Abundant is an applied research lab focused on scaling reinforcement learning for safe and reliable agentic capabilities. We are an extremely talent-dense team of researchers, roboticists, founders, and operators whose work includes the Waymo Driver.
THE ROLE
Frontier labs test their models in environments like ours, which means we are the first to see novel model behavior. You will make sure that alignment is baked into everything the do: the code we deploy, the runbooks we write, and the papers we publish.
WHAT YOU’LL DO
Show full descriptionShow less
Run experiments on how our environments shape behavior: reward hacking, cheating the task, unsafe use of tools, deception over long runs. Then change how we build them.
Break our own graders. Find the ways to score well without doing the work, close them, and keep the attacks as tests everyone runs.
Build the tools that watch agents: monitors over trajectories, red team harnesses, and safety benchmarks that stay hard as models improve. On real runs, not toy ones.
Keep our environments sealed. Network isolation, escapes, credentials, and how much damage an agent can do when it turns on us. Last summer’s failures came from here.
Choose a research question about oversight, control, or long-horizon autonomy, answer it, and publish. This is part of the job, not something you do at night.
Decide what we will not build. Tell partners plainly what they get from us is and is not safe for.
WHO YOU ARE
You are an engineer first. You build your own test pipelines, learn strange codebases quickly, and find the bug in the log.
You know LLM safety in depth: attacks, red teaming, monitoring and control evaluations, reward hacking.
You can turn a vague worry about a model into an experiment, run it, and say what the result does and does not prove.
You read transcripts closely. You catch the small thing that makes the whole result wrong.
You have put research into production systems, especially post-training, distillation, or evaluation that something depends on.
You write clearly enough that your results change what other people do.
You can decide fast with incomplete evidence and defend the call with data.
NICE TO HAVE
Published work on control, dangerous capability evaluations, oversight, or interpretability.
RLHF or RLAIF experience, and a view on how training choices show up in behavior.
You built or ran a public benchmark or agent task suite.
You have run many agents at once: sandboxes, containers, logging.
You have advised on AI safety or governance policy.
COMPENSATION
Base Salary
$250,000 - $450,000
Cash Bonus
Sizable performance bonus tied to project and company milestones