Member of Technical Staff, AI Products (Early Career - Industry)
PerplexitySan Francisco, CAposted 22d ago
- PayUSD / yr
- $180k – 270k
- Where
- San Francisco, CA
- Posted
- Sep 23
- 22 days ago
- Type
- Full-time
- Setup
- On-site
At a glance
Build and ship full-stack, AI-powered product experiences across Perplexity’s search, agentic, and model-driven surfaces, taking ideas from prototype to production and improving them with real-world data and evals.
Summarized by AI from the original posting
What you'll do
- Own a user-facing surface end to end across interaction, design, frontend, application logic, APIs, data models, and search or model calls
- Ship continuously to production
- Use AI agents and coding models effectively
- Prototype ambiguous product ideas and evaluate them using usage data and evals
- Improve reliability, latency, and quality
- Collaborate with designers, product managers, and research engineers
Requirements
- 01Production software engineering experience
- 02Strong backend language proficiency in Python or Go
- 03Experience with services, APIs, and relational data modeling
- 04Frontend experience with React and TypeScript
- 05Hands-on experience with AI products
- 06Comfort with ambiguity and incremental development
- 07Experience building on top of LLMs preferred
- 08Public track record such as GitHub, technical blog, or demo preferred
Skills
Full description
NOTE ON ELIGIBILITY: this role is for industry candidates already working in software engineering for the past 1+ years. New grads and current students should NOT apply to this role (instead, visit pplx.ai/newgrads).
About the role
In 2026, we launched Computer, the defining product for the new era of agentic AI. Millions of people now use Perplexity to transform knowledge into action, through Search, Comet, and Computer. Product Engineering owns those experiences end to end, from the interface a user touches to the search and model calls behind it, on top of roughly 200M queries a day and a 200B+ URL index.
This is a full stack role in the truest sense. There are no frontend or backend lanes here, and no one waits for a ticket before starting. It suits people who love to build and want the room to do it. What you have shipped and how quickly you learn tell us far more than years on a resume or where you went to school.
What you will do
Own a user-facing surface end to end, from the interaction, design, and frontend through the application logic, APIs, data models, and the search or model calls behind them.
Ship to production continuously. New engineers typically have code in front of users in their first week.
Build with AI as a primary tool, not a novelty. We expect you to use agents and coding models aggressively to move faster, and to have sound judgment about when their output is wrong.
Turn ambiguous product ideas into working prototypes, then use real usage data and evals to decide what survives.
Raise the quality bar on reliability, latency, and craft for the surfaces you own.
Work directly with designers, PMs, and research engineers, without unnecessary process in between.
What we look for
1+ years of professional software engineering experience, or a track record of strong internships where you shipped production code at scale.
Strong in at least one backend language, such as Python or Go, with real experience across services, APIs, and relational data modeling. Comfortable working in React and TypeScript on the frontend.
Real evidence of building: production work you understand in depth, shipped side projects with actual users, meaningful open-source contributions, hackathon wins, or research you turned into working software.
Comfort with ambiguity and an instinct to start building incrementally rather than wait for a spec.
Hands-on experience with AI products, with opinions about their rough edges and what that means for the product you are building.
Nice to have
Experience with our stack: React or Next.js, Python, Go, PostgreSQL, Redis, Docker, AWS. You will not be expected to know all of it on day one.
Experience building on top of LLMs: AI orchestration, RAG, evals, agents, tool use.
Product taste. You can tell a good interface from a bad one and explain how to improve it.
A public track record we can read, such as GitHub, a technical blog, or a demo we can try.