Applied AI Research Engineer

Applied Research Engineer Frontier AI Systems
Compensation: $200,000 - $300,000 USD (plus equity)
Location: San Francisco preferred | Hybrid (2 days/week in-office)

Who are we?

Were building the infrastructure and expertise that powers the next generation of AI models. Our platform bridges advanced machine learning with high-quality, domain-specific data to solve some of the most complex challenges in frontier model development. From fine-tuning LLMs with expert human feedback to optimizing large-scale expert networks across disciplines like physics, linguistics, and lawthis is where cutting-edge research meets real-world impact.

We partner directly with leading AI labs to create the specialized, PhD-level data their models require. If we succeed, well redefine how AI systems learn, align with human values, and reason across domains.

What's in it for you?

• Real Authority, Day One: You'll step into a senior IC role with high autonomy, joining a team that already collaborates with frontier labs.

• Mission-Critical Impact: You'll design systems and methods that shape the data and feedback loops used to train the worlds most advanced LLMs.

• Startup Jungle, Research Depth: Operate at speed in a startup environment that still values deep thinking and first-principles innovation.

• Frontier-Level Challenges: No prebuilt tools, no plug-and-play. Youll invent new ways to assess data quality, optimize expert selection, and scale human feedback with precision.

What will you do?

• Develop novel techniques for AI-human alignment, including RLHF, DPO, and other human-in-the-loop training strategies.

• Build and scale systems to assess and optimize expert networks, assigning the right annotators to the right PhD-level tasks.

• Create evaluators to score expert competence and data complexity, such as automated review systems for math proofs or legal analyses.

• Design active learning and adaptive sampling tools that reduce manual labeling without compromising quality.

• Prototype AI-assisted interview systems to assess domain knowledge and task suitability at scale.

• Publish and present research that shapes how the industry approaches data-centric AI, from NeurIPS to EMNLP.

• Collaborate closely with frontier research labs, translating their needs into practical, scalable tooling.

What will you need?

• 3+ years shipping impactful work with LLMs or related systems fine-tuning, pretraining, evaluation, or large-scale deployment.

• A portfolio of work demonstrating battle-tested insights from the bleeding edge of ML (your scars are welcome).

• Experience designing and shipping applied ML systems that balance rigor with real-world delivery.

• Strong foundation in Python and frameworks like PyTorch, JAX, or TensorFlow.

• Deep curiosity and technical fluency across multiple domains (AI alignment, optimization, domain-specific modeling, etc.).

• Graduate-level training in CS, ML, or related field (PhD or MS preferred) and a solid publication track record.

• High agency mindset you're not waiting to be told what to build.

• Ability to operate in high-context, low-structure environments (think: startup jungle with frontier labs as your neighbors).

Not a fit if...

• You've only worked on toy LLM applications or inside rigid big-tech environments.

• You're married to a specific subfield and allergic to ambiguity.

• You're more interested in polishing code than solving hard, undefined problems.

This is an urgent hire. Were looking for someone who can hit the ground sprinting, contribute with authority, and isn't afraid to build where no playbook exists.

If you've shipped something significant with LLMs and are ready to help architect the future of AI, we want to hear from you.