Latest AI JobsBrowse Jobs
A

Member of Technical Staff, Alignment

abundantSan FranciscoEngineering
Apply on company site
Salary
$250K - $450K
Job type
Full-time
Experience
Senior
Posted
16 Sept 2026
Status
Open · checked 7d ago

About this role

San Francisco is the location for the Member of Technical Staff, Alignment position at Abundant, where an engineer drives safety and control protocols across applied reinforcement learning research. This role suits professionals with deep expertise in large language model safety, agent oversight, and production-level machine learning systems.

Role Overview

Abundant operates as an applied research lab scaling reinforcement learning for safe agentic capabilities. The organization builds environments where frontier labs test models, placing the team at the forefront of observing novel model behavior. Staff in this position ensure alignment remains integrated into deployed code, published papers, and operational runbooks.

Key Responsibilities

Conduct experiments regarding environment impact on behavior, including reward hacking, task cheating, tool misuse, and long-run deception.
Identify and eliminate vulnerabilities in grading systems while turning attacks into standard automated tests.
Develop monitoring instruments for agent trajectories, red team frameworks, and scalable safety benchmarks on real runs.
Maintain environment security through network isolation, credential management, and damage limitation.
Pursue dedicated research questions concerning oversight, control, or long-horizon autonomy and publish the findings.
Determine product boundaries and clearly communicate safety parameters to partners.

Requirements

Function primarily as an engineer who builds test pipelines, navigates codebases rapidly, and diagnoses log errors.
Possess thorough knowledge of large language model safety concepts such as attacks, red teaming, evaluations, and reward hacking.
Translate abstract concerns regarding model behavior into structured, verifiable experiments.
Analyze transcripts meticulously to catch discrepancies in outcomes.
Deploy research directly into production systems, focusing on areas like post-training, distillation, or evaluation.
Communicate research conclusions clearly to influence team actions.
Make rapid decisions using partial evidence and support conclusions with data.

Nice to Have

Published literature concerning control, dangerous capability evaluations, oversight, or interpretability.
Practical background in RLHF or RLAIF alongside perspectives on training choices.
Experience creating or operating public benchmarks or agent task suites.
Background running concurrent agents utilizing sandboxes, containers, and logging.
Advisory history regarding AI safety or governance policy.

Skills and Tools

Large Language Models
Reinforcement Learning
RLHF
RLAIF
Python

Salary and Benefits

Base salary ranging from $250,000 to $450,000.
Sizable performance bonus linked to project and company milestones.
Generous early-stage equity grant.
Health, dental, and vision insurance alongside flexible paid time off.

About the Company

Abundant is an applied research lab focused on scaling reinforcement learning for safe and reliable agentic capabilities. The team consists of researchers, roboticists, founders, and operators with experience that includes the Waymo Driver.

How to Apply

Interested candidates should apply on the official company careers page.

Skills and tech stack

LLMRLRLHFRLAIFReinforcement Learning

Questions candidates ask

Where is the job located?

The position is located in San Francisco.

What is the salary range for this role?

The base salary is between $250,000 and $450,000, supplemented by a performance bonus and early-stage equity.

What background is required for this position?

Candidates must be engineers with deep knowledge of LLM safety, red teaming, reward hacking, and experience deploying research into production systems.

Is research publication part of the job?

Yes, choosing a research question about oversight, control, or long-horizon autonomy and publishing the results is a formal part of the position.

Ready to apply?

The application is on the official abundant careers page.

Apply on company site