Head of Research (AI/ML)

Ambral Labs; San Francisco (CA); On-site; USD 180,000 - 290,000; Full time

Job summary

Ambral Labs seeks a Head of Research to lead the agenda for building a replayable environment engine over real enterprise history. You will define high‑leverage questions, design robust experiments, and stay hands‑on in turning research into production systems.

You will collaborate with the CTO and scale research across thousands of concurrent runs, advancing post‑training methods for agents operating over long horizons and large toolspaces.

Qualifications

• You have a PhD in machine learning, computer science, mathematics, or an equivalent track record of significant research experience
• You have deep experience in reinforcement learning, LLM post‑training, evals, agent environments, or closely related areas
• You've taken ambitious, open‑ended research problems from hypothesis through experimentation into working systems
• You’re comfortable turning fuzzy business objectives into tasks and signals that can be evaluated reliably
• You can move between research questions and production implementation without treating them as separate jobs
• You’re looking to do the best work of your life and build something you’ll be proud of for decades
• We’re especially interested in candidates who have worked at a leading foundation model lab, top AI research organization, or high‑performing AI startup.

Responsibilities

• Own the research agenda required to make replayable environments possible
• Identify the highest‑leverage technical questions and design experiments to answer them
• Remain deeply hands‑on in building the systems that turn answers into production
• Contribute to defining how Ambral runs research and helps grow an exceptional team
• Collaborate with CTO and production teams to scale research to thousands of runs

Skills

Reinforcement learning

LLM post-training

Experiment design

Research leadership

Education

PhD in machine learning / CS / mathematics

Job description

Ambral Labs helps enterprises own the intelligence behind their most important workflows.

Every company has years of historical evidence showing how work gets done: the context people had, the decisions they made, the actions they took, and the outcomes that followed. Today, most of that history is inert. It isn’t structured in a way that companies can use to evaluate models and improve agent behavior.

Ambral turns this history into replayable environments and eval sets grounded in real workflows and observed outcomes. We use those environments to help improve performance of open-weight through reinforcement learning and other post-training techniques, alongside context engineering, harness design, and agent engineering. The result is better, more cost-efficient AI for each enterprise’s specific work, powered by open-weight models that the company can own and control rather than permanently renting from a model provider.

We're a YC S2025 company, have raised millions in funding, and are already deployed inside multi-billion dollar enterprises. Now we're growing the founding team.

What you'll do:

We’re building a replayable environment engine over real enterprise history.

The system reconstructs a company’s context as it existed at any past time, then exposes that state through the same tools an agent would use in production. This lets us place new policies and agent configurations inside real historical environments, observe how they reason and act, and grade their performance against real outcomes.

As Head of Research, you'll own the research agenda required to make that possible. You'll identify the highest-leverage technical questions, design the experiments needed to answer them, and remain deeply hands‑on in building the systems that turn those answers into production.

Some of the problems you'll work on:

• Building an environment factory that converts recorded enterprise data and task definitions into runnable environments


• Designing graders that turn ambiguous business objectives into verifiable rewards


• Developing methods for mining useful tasks, trajectories, and evaluation cases from historical workflows


• Creating eval sets that are representative, reproducible, and resistant to overfitting


• Finding the right combinations of models, tools, context, and policies to maximize performance while reducing inference cost


• Advancing post‑training methods for agents that operate over long horizons, incomplete information, and large tool spaces


• Building replay and observability systems that make agent behavior explainable and measurable


• Scaling from individual environments to thousands of concurrent training and evaluation runs

These problems are wide open. You’ll work directly with the CTO, own major parts of the research direction and production systems, and see your work tested against consequential problems in real enterprise workflows. You’ll also help establish how Ambral runs research and, over time, build an exceptional research team.

What we do:

You'll likely thrive here if:

• You have a PhD in machine learning, computer science, mathematics, or an equivalent track record of significant research experience


• You have deep experience in reinforcement learning, LLM post‑training, evals, agent environments, or closely related areas


• You've taken ambitious, open‑ended research problems from hypothesis through experimentation into working systems


• You’re comfortable turning fuzzy business objectives into tasks and signals that can be evaluated reliably


• You can move between research questions and production implementation without treating them as separate jobs


• You’re looking to do the best work of your life and build something you’ll be proud of for decades

We’re especially interested in candidates who have worked at a leading foundation model lab, top AI research organization, or high‑performing AI startup.

Compensation: