Principal AI Engineer

Principal AI/ML Engineer

Full Time Position

Salary: $350,000 USD

Equity benefits

100% Remote

Location & Logistics

• Ideally candidates located in Seattle or Portland.

The Ro

leWe’re hiring a SENIOR AI/ML Engineer for a research-heavy R&D role focused on Machine Learning, specifically Reinforcement Learning (RL) in an LLM context. This role is not defined by degrees or pedigree, we value proof-of-work, tenacity, and the ability to solve exceptionally hard problems for a company that is transforming how content is created and trusted in publishing. They build technology that is accurate, scalable, and creative—designed to elevate both craft and integrit

y.


Typical monthly time

• split20% res
• earch40% model tra
• ining30% performance optimiz
• ation10% production

work
You will primarily fine-tune existing open-weight foundation models (e.g., Llama, DeepSeek) based on research, innovation, and experimental findings. We use a custom RL method, and you’ll be expected to implement, experiment, improve, and innovate on it (not design it from scr

atch).
What Yo

• u’ll DoImplement and improve reinforcement learning algorithms used in model t
• rainingPartner closely with the data team to define reward signals and training obj
• ectivesFine-tune and refine models using open-weight and proprietary foun
• dationsRun experiments, analyze results, and iterate based on empirical o
• utcomesTurn research ideas into stable, repeatable training pi
• pelinesDebug training instability and improve training performance (efficiency, reliability, thro

ughput)
Required Qualifications (Non-Nego

• tiables)Hands-on experience with Reinforcement Learn
• ing (RL)Hands-on experience with LLMs (training or fine-tuning in real systems, not just
• theory)Strong Pytho
• n skillsStrong PyTorch ex
• perienceStrong R&D / experimentation skills (designing experiments, analyzing outcomes, it
• erating)Excellent technical documentation ability (clear experiment writeups, white papers, desig
• n notes)Ability to work with scoped requirements, while applying creative problem
• -solvingAbility to articulate work clearly both orally and in writing to technical and non-technical a

udiences
Nice-to-Have (Str

• ong Plus)CUDA familiarity and/or performance optimization e
• xperienceWe do not expect you to write custom CUDA kernels, but we value the ability to debug/optimize training pe
• rformanceDistributed training / large-scale experimentation e
• xperiencePrior work on LLM training stacks or modern fine-tuning
• workflowsPublished research, patents, or strong proof-of-work
• artifactsStartup / fast-moving environment e

xperience
H

• ow we workHighly collaborative environment: we learn from each other and make decision
• s togetherYou’ll work from scoped priorities, but we value initiative, clarity, and creative
• iterationInfrastructure/tooling specifics may vary; you’ll succeed by being adaptable and