About Aaru
Aaru builds simulations of human behavior. Each simulation contains a population of AI agents, each representing a person who could plausibly exist in the real world and capable of making decisions within a modeled environment. Companies and institutions use these simulations to test consequential choices before committing—from product launches and pricing decisions to strategic communications and policy changes.
Building a useful simulation requires more than generating plausible text. Populations must represent real people and groups; predictions must be calibrated; simulations must remain coherent as conditions change; and the product must make the resulting evidence legible enough to support real decisions.
We are a small, in-person team in New York. We work with urgency, high ownership, and intellectual honesty. We expect people to surface inconvenient evidence, change their minds quickly, and carry important work all the way to a result.
About Product Engineering
Product Engineering turns Aaru's simulation capabilities into products that customers can use independently and repeatedly. The team builds on Aaru's shared platform and simulation systems to create the workflows, interfaces, integrations, and decision-ready artifacts that make a technically sophisticated system feel clear and dependable.
The work is organized around durable product domains rather than a queue of isolated features. Those domains may include simulation setup, question and experiment types, follow-up and continuous simulations, analysis and reporting, reusable customer templates, collaboration, and external integrations.
This is not a thin frontend role. Product Engineers own user outcomes across the stack. A project may begin with observing a customer's decision process, continue through product and systems design, require new backend or orchestration primitives, involve careful evaluation of model behavior, and end with a measured production rollout.
The role
As a Product Engineer, you will take ambiguous, consequential user problems from first conversation to reliable product capability. You will work closely with Product, Design, Deployment, Platform Engineering, Simulation Engineering, and Research. You will be expected to understand both the user problem and the system details well enough to make good tradeoffs without handing ownership away at either boundary.
You will prototype quickly, but you will not confuse a compelling demo with a finished product. The systems you ship must be understandable, observable, secure, maintainable, and robust to the variability of AI-generated behavior. When a shared primitive is missing, you will work through the platform boundary or help create it rather than building a fragile one-off around it.
What you will do
Own product work end to end: understand the problem, define the smallest useful solution, design the system, implement it, roll it out, measure it, and support it in production.
Work directly with users and customer-facing teams to observe real decision workflows, identify recurring needs, and distinguish durable product opportunities from bespoke requests.
Build polished customer-facing experiences across frontend, backend, APIs, data models, workflow orchestration, permissions, integrations, and AI-driven interactions.
Translate capabilities from Population Research, Prediction Research, Evaluation Research, and Simulation Engineering into product experiences that customers can use without expert assistance.
Design product behavior around the uncertainty and variability of AI systems, including clear states, human review points, fallbacks, retries, provenance, and honest communication of confidence.
Define evaluation and launch criteria before shipping model-dependent features. Use offline evaluations, product signals, operational metrics, and qualitative feedback to determine whether a change is actually better.
Turn specific customer evidence into generalizable product primitives, templates, and workflows rather than accumulating one-off branches and configuration.
Work with Platform Engineering through clear interfaces, contribute missing primitives when appropriate, and avoid coupling product delivery to undocumented platform behavior.
Instrument adoption, task completion, quality, latency, cost, reliability, and failure modes so that product decisions are based on evidence rather than anecdotes.
Own the operational quality of what you ship, including production support, debugging, incident follow-up, migrations, and safe rollback paths.
Write clear technical designs, product notes, and launch documentation. Make scope, assumptions, dependencies, and unresolved risks legible to the rest of the company.
Raise the quality bar through thoughtful code review, testing, design critique, and improvements to the tools and patterns used by the broader engineering team.
Representative problems
You might work on problems such as:
Add a new question or allocation format that requires changes to the product interface, simulation contract, validation logic, analysis layer, and customer-facing output.
Build a continuous simulation product that ingests new information over time, updates relevant assumptions, and shows users what changed and why.
Create a follow-up workflow that lets a user interrogate a completed simulation without losing provenance, population state, or the distinction between observed and generated evidence.
Turn a customer's successful simulation setup into a reusable organization-specific template with sensible defaults, permissions, versioning, and audit history.
Build an agent-assisted setup experience that helps users specify a decision, identify missing context, and construct a valid simulation without hiding important assumptions.
Create decision-ready reports, presentations, or interactive artifacts that preserve uncertainty and trace each conclusion back to the underlying simulation evidence.
Integrate Aaru with a customer's data or operating system while handling authentication, permissions, data mapping, validation, retries, and safe writes.
Diagnose a feature whose demo looks excellent but whose real-world completion rate is poor, determine whether the failure is in the model, workflow, interface, or expectation setting, and ship the right fix.
How we work
We begin with the decision the user is trying to make, not with a feature request. We seek direct evidence, reduce the problem to its essential uncertainty, and build the smallest system that can resolve it. We move quickly, but we preserve the foundations required for reuse and long-term ownership.
For AI-native products, product quality and model quality are inseparable. Latency, interaction design, evaluation, orchestration, permissions, reliability, and the behavior of the underlying models all shape the user experience. Product Engineers are expected to reason across these layers rather than treating model behavior as somebody else's API.
We prefer clear ownership and small teams. The person closest to a problem should have the context and authority to make decisions, while documenting enough that the rest of the organization can understand and challenge them.
You might thrive in this role if
You have built and operated meaningful user-facing products, ideally from an early or ambiguous starting point.
You combine product judgment with strong software engineering fundamentals and can move comfortably between user workflows and system architecture.
You can work across a modern product stack and are willing to learn whichever layer is necessary to finish the job.
You seek out users, ask precise questions, and can tell the difference between what someone requests and what would actually solve their problem.
You have strong taste for simple workflows, clear interfaces, and the small details that make a product feel fast, trustworthy, and coherent.
You make sensible tradeoffs between iteration speed, quality, scope, maintainability, and technical risk—and can explain those tradeoffs plainly.
You treat instrumentation, evaluation, and production support as part of building the feature rather than work to be added later.
You are comfortable with loosely specified problems, shifting information, and situations where the right answer must be discovered through prototypes and evidence.
You take responsibility for outcomes, including when the problem crosses team boundaries or the first approach fails.
You want to work in person in New York with a team that moves quickly and debates the work directly.
Strong candidates may also have
Experience as a founder, founding engineer, or early product engineer at a fast-growing company.
Experience building AI-native products, agentic workflows, copilots, or products in which model behavior materially shaped the user experience.
Fluency with technologies such as TypeScript, React, Python, APIs, workflow systems, relational data models, and cloud infrastructure—or equivalent depth in a comparable stack.
Experience with B2B or enterprise products involving permissions, auditability, integrations, collaboration, or complex configuration.
Experience with data visualization, analytical products, experiment builders, survey tools, research platforms, or decision-support systems.
A record of turning a customer-specific need into a reusable product capability adopted by many users.
Experience working closely with research or machine-learning teams and translating experimental capabilities into dependable product behavior.
Candidates need not have
Prior experience in simulation, computational social science, or Aaru's exact technical stack.
A conventional computer-science degree or a specific number of years in a particular title.
Every skill listed above. We care most about evidence that you can understand hard user problems, build excellent systems, and learn quickly.
What success looks like
Customers can configure, run, and interpret important simulations with substantially less expert assistance.
New research and simulation capabilities become understandable, reusable product features rather than bespoke demonstrations.
The product team ships quickly without accumulating fragile workflows, hidden assumptions, or avoidable platform debt.
Product quality is measurable: adoption, task success, model behavior, latency, reliability, cost, and failure modes are visible and improve over time.
Specific customer learnings become durable product patterns that compound across accounts and use cases.
The systems you own are reliable in production, easy to debug, and clear enough for other engineers to extend.
Colleagues trust you to take an important, ambiguous product problem and carry it through to a high-quality outcome.
Location and benefits
This role is based in New York City. Aaru is an in-person company, working five days a week in the office. Candidates should be located in the New York metropolitan area or open to relocation.
Aaru offers a competitive base salary, equity participation, comprehensive medical, vision, and dental coverage, visa sponsorship and relocation support, and other benefits and perks. Final compensation depends on level and experience and is set within Aaru's internal bands.