Member of Technical Staff (Data & RL Environments)

About Pavo

Pavo is building the Applied Science Factory: a system that develops an understanding of production systems, investigates performance, constructs reliable evaluations, and helps engineers improve business outcomes.

We work with the underlying product, code, data, models, and experiment history. Our research spans training environments, scientific agent capabilities, and models for structured and behavioral data.

We're a small team with experience building large-scale ML systems at Spotify, ShareChat, and Sourcegraph. Members of Technical Staff own research and implementation end to end.

The role

Pavo's agents need to learn how to investigate production systems using code, data, product behavior, and experiment results. You'll develop the datasets and executable environments used to train and evaluate these capabilities.

A representative environment could simulate users, products, transactions, recommendation decisions, experiment assignment, and downstream metrics. Agents would query the data, inspect decisions, run analyses, and test hypotheses through a controlled set of tools.

You'll own environment design, data generation, task construction, tool interfaces, verifiers, and the pipeline for collecting and evaluating agent trajectories.

What you'll build

• Build stateful simulations of user journeys, transactions, ranking or pricing policies, experiments, and delayed outcomes, with consistent data and metrics.
• Create applied-science tasks such as diagnosing a metric regression, identifying cohort-level effects, or constructing an offline evaluation.
• Implement sandboxed tools, environment reset and replay, trajectory capture, programmatic checks, and statistical verifiers.
• Develop pipelines for synthetic data, expert demonstrations, agent rollouts, quality filtering, and held-out environment splits.
• Evaluate task difficulty, leakage, reward hacking, and transfer from simulated environments to unfamiliar problems.

What success looks like

• A reproducible executable environment with coherent data, working tools, and validated tasks and verifiers.
• Agent rollout analysis and a usable pipeline for generating training and evaluation data.

What we look for

• Have built substantial simulation, synthetic-data, agent-environment, or evaluation infrastructure.
• Strong Python and systems engineering; comfortable with sandboxed execution, data pipelines, and agent tool interfaces.
• Sound grounding in statistics and experimental design, including randomized policies, delayed feedback, and observational data.
• Experience designing evaluations or verifiers, inspecting model trajectories, and debugging unexpected behavior.

Relevant experience

• RL environments, agent rollouts, recommendation systems, marketplaces, causal simulation, or experimentation platforms.

What we offer

• Own a substantial area of research and engineering, working directly with the founders.
• Work on production ML problems and develop reusable methods and systems. Salary and equity.

Pavo is an equal opportunity employer.