About the job Applied AI Research Engineer
Applied AI Research Engineer
Compensation: $250,000$350,000 USD
Location: San Francisco, Hybrid (2 days/week onsite)
Who Are We?
We're a fast-moving AI infrastructure company partnering with frontier labs and top-tier research organizations to deliver the high-quality, domain-specific training data that cutting-edge models depend on. Our platform and services are critical for teams working on advanced AI systems from LLMs to multimodal model and our mission is to push the frontier of what's possible with human-in-the-loop AI development.
Were not here to follow trends were here to set them. Think of us as the silent engine behind your favorite frontier AI breakthroughs. You may not know our name, but you've definitely seen our fingerprints on today's most advanced models.
What's In It for You?
-
Real Impact, Real Fast: You'll be dropped straight into a jungle not a garden and expected to thrive. Immediate ownership, high autonomy, and direct access to some of the most sophisticated AI teams on the planet.
-
A Seat at the Big Kids Table: Collaborate with ML engineers, researchers, and AI leads from elite frontier labs. Your work will directly shape how next-gen models are trained and aligned.
-
No Off-the-Shelf Problems: You'll tackle gnarly, unsolved challenges like evaluating PhD-level expert networks or automating assessments of LLM-generated content. If you're looking for boilerplate ML, this isn't it.
-
Continuous Learning: You'll be in an environment where curiosity isn't just encouraged its required. We prize intellectual sharpness and a bias for execution over credentials.
-
Pay That Reflects Your Scar Tissue: We're looking for people who've been through the LLM trenches and made it out smarter. You'll be compensated accordingly.
What Will You Do?
-
Build systems that align human feedback into training loops think RLHF, DPO, and methods not even named yet.
-
Design algorithms to automatically evaluate the quality of human feedback and data at scale.
-
Prototype tools to support expert labeling and assessment across domains like physics, math, linguistics, and more.
-
Develop automated systems to assess expert competency and route the right task to the right human (yes, kind of like dating apps, but for PhDs and LLMs).
-
Research and implement active learning strategies, adaptive sampling, and other methods to minimize human effort while maximizing model performance.
-
Publish in top-tier ML/AI conferences and build thought leadership in AI alignment and human-in-the-loop training.
-
Collaborate closely with top-tier customers (think: frontier labs and household tech names) to understand real-world model training needs and translate them into scalable systems.
What Will You Need?
-
Experience shipping significant projects in LLMs or adjacent areas pretraining, post-training, fine-tuning, evaluation, human alignment you've done it, and you've got the scars to prove it.
-
A proven ability to walk into complex ML problems and start delivering value fastideally in environments where startup chaos was a polite understatement.
-
Strong grounding in ML theory and applied research, ideally backed by a PhD or Masters in Computer Science, Machine Learning, or related field.
-
Fluency in Python and experience with deep learning frameworks like PyTorch, JAX, or TensorFlow.
-
A track record of publication in top-tier AI/ML conferences (NeurIPS, ICML, ICLR, ACL, etc.) is a big plus.
-
High agency, high integrity, low ego.
-
Bonus: Experience in assessing domain expertise, automated evaluation, or optimization problems involving humans-in-the-loop.