Product Engineer: AI-Powered, Customer-Driven IC

Most engineering jobs start with a ticket. This one starts with a customer problem.

At Metaview, you will talk directly to the people using what you build. You will decide what should exist, build the consequential parts across the stack, put it in front of customers, and stay responsible for what happens next.

There is no chain of handoffs between product, design, engineering, support, and the customer. The person closest to the problem owns the result.

Metaview is an applied AI lab building AI coworkers for some of the world’s most ambitious recruiting teams. Hear directly from our customers. The challenge is not another chatbot. It is software that can pursue a hiring objective, act across real workflows, ask for judgment at the right moment, and earn a customer’s trust.

Founded by Siadhal Magos and Shahriar Tajbakhsh, who scaled Uber and Palantir, we have raised more than $50M and are growing 5x year over year.

AI coding agents are part of the job, not a perk. Engineers routinely run several in parallel for exploration, implementation, and review. We assess both sides of the craft: how well you direct agents and how well you reason, code, and debug on your own. One interview exercise is completed without AI.

Velocity means reducing the time from customer evidence to a reliable product change. It does not mean trading away quality. We ship small changes daily, including Fridays, because the engineer who ships owns production behavior, recovery, and the next revision. Process stays light because customer context and ownership are direct.

No one will hand you a polished specification. You will be expected to understand the customer, form an opinion, choose a path, and get working software into use. You will also own quality, production behavior, and the revision that follows the first release.

AI agents help our small team ship roughly 8x more product changes than a year ago. The engineer remains accountable for architecture, review, implementation quality, and production behavior.

You will work on problems such as:

• building agents that can pursue recruiting objectives over days, know when to ask for judgment, and recover safely when a model or tool fails;

• turning interview and hiring-team conversations into actions, with interfaces that let recruiters understand, correct, and trust the result;

• designing permissions, controls, and recovery paths for agents acting in customer systems;

• building the interface, APIs, data model, and production behavior of the same customer workflow; and

• turning repeated product patterns into simple primitives that support many recruiting workflows.

In your first 3 to 6 months, you should independently own and ship a bounded customer outcome from problem definition through rollout and iteration. The size of that outcome will reflect your level, but the operating model is the same: make product and technical decisions, contribute the consequential code, and own the production result.

• You have personally shipped software for external users and owned a meaningful outcome from problem definition through production.

• You have real depth in at least one technical area and can cross product, frontend, and backend boundaries when the outcome requires it.

• You use AI coding agents heavily and can still design, implement, review, and debug independently.

• You want direct customer contact, hands-on implementation, and responsibility after launch.

• You naturally start with the customer problem rather than the requested feature.

• You have strong product opinions and change them when evidence proves you wrong.

This opening spans mid-level and staff+ scope under the same hands-on operating model. At mid-level, you will independently own a bounded customer outcome across product and implementation. At staff+, you will still write code while setting direction for a broader area, making hard architectural tradeoffs, and raising the quality of decisions around you. We level on demonstrated scope, not years served.

This is a hands-on product IC role. It is not a people-management or model-research role.

Our stack includes React, Type Script, Python, AWS, Postgres, and S3. Experience with these exact technologies is not required.

• We work together in person in San Francisco. This is not a remote role with occasional office attendance. Decisions are made by the people…