Perplexity

perplexity.ai · 110 open roles in the last 6 months · hiring in 5 countries

Loading…

Perplexity appears to be investing in agentic products—especially Computer—and the enterprise controls, evaluation systems, and customer-facing teams needed to make them useful at work.

Inferred from 126 job postings from the last year, with recent ones weighted more · updated 21 Oct 2026. Not a statement by the company.

What changed recently

The last month adds a sharper enterprise-readiness layer: staff roles target agent evaluation, workflow reliability, identity, permissions, and governance, alongside new enterprise sales leadership. It also extends the consumer push for Computer through iOS and Android growth work, while dedicated API customer-success roles from the prior few months point to a more hands-on developer adoption motion.

Scalinghigh confidenceMostly new this month

Improving search quality across crawling, ranking, and agent use

1m 53m 86m 81y 8

Search remains a substantial technical investment, spanning URL discovery and parsing through indexing, retrieval, ranking, and agent search. Postings cite an index of more than 200 billion URLs and systems processing billions of pages; a recent Search Agent RL role adds a focus on post-training, multi-agent systems, and better task completion.

Recent roles address both crawling quality and machine-learning improvements to search agents.
The work covers retrieval and ranking models, indexing pipelines, RAG, and search infrastructure.
A product engineering posting cites roughly 200 million queries a day and a 200-billion-plus URL index.

Scalinghigh confidence

Growing Computer through mobile adoption and paid conversion

1m 23m 66m 81y 8

Computer is being treated as a major product surface, with iOS and Android engineers focused on activation, retention, and paid conversion across Perplexity, Comet, and Computer. Product roles also describe work on skills, collaboration, artifact generation, onboarding, and upgrade flows, indicating investment in making agentic capabilities easier to adopt and return to.

Two recent mobile engineering roles focus on growth experiences across Computer, Comet, and Perplexity.
Product work spans activation, repeat usage, paid conversion, collaboration, and agent capabilities.
The postings describe products used by millions of people.

Scalinghigh confidenceMostly new this month

Making Computer and agents work securely inside enterprises

1m 33m 56m 71y 7

Perplexity is building enterprise capabilities around Computer and agents, from the control plane and connector marketplace to identity, provisioning, permissions, and admin APIs. The recent focus on enterprise agent evaluations and governance suggests an effort to help organizations move from pilots to broader rollouts, with agents completing workflows reliably and with enforceable access controls.

Recent staff roles separately target enterprise agent evaluation and trust/governance.
Postings describe an enterprise control plane, data plane, connector marketplace, and knowledge platform.
Enterprise workflows include Slack, Microsoft Teams, and Microsoft Office, with human approvals.

Scalingmedium confidence

Reducing agent and inference costs while improving reliability

1m 13m 56m 51y 7

Perplexity is investing in the infrastructure and measurement needed to serve AI products at scale: model serving, GPU compute, inference, and cross-cutting retrieval and serving systems. Recent enterprise evaluation work explicitly targets lower latency and cost per successful task, while an earlier inference-platform posting describes hundreds of millions of model requests per day and support for a changing portfolio of models.

Agent evaluation is tied to reliability, latency, and cost per completed task.
Postings cover GPU capacity planning, inference serving, model adoption, and infrastructure efficiency.
The model-serving platform posting cites hundreds of millions of requests per day.

Scalinghigh confidence

Expanding the API platform and helping customers put it to work

1m 03m 46m 41y 4

The API platform packages Perplexity’s search, AI, and agent orchestration capabilities for developers, with a posting describing thousands of customers ranging from startups to large enterprises and governments. API engineering is paired with newer customer-success and business-operations roles focused on onboarding, adoption, retention, and moving integrations from evaluation into production.

Recent API roles add customer success and post-sale operations to platform engineering and sales.
The platform includes Search, Sonar, and Agent APIs.
The postings describe customers from AI-native startups to large technology companies.

Where they are expanding

London — A dedicated Engineering Site Lead role describes establishing a London engineering organization, including infrastructure and AI-infrastructure teams, with an expected 15–30-plus engineers after a year. Recruiting and technical roles in London and nearby European locations reinforce this as more than a single isolated opening.

Questions to ask in a coffee chat

How are the enterprise agent evaluation and governance teams dividing responsibility for moving Computer deployments from pilots to wider rollouts?
Which parts of Computer’s mobile experience are the team prioritizing most right now: onboarding, repeat use, or paid upgrades?
How does feedback from API customers move from customer-success conversations into platform or product changes?
What are the hardest search-quality problems the team is tackling across crawling, indexing, ranking, and agent search?
How is the London engineering site expected to collaborate with the existing Search, AI, and infrastructure teams?

What this can't tell you

Job postings show areas of active hiring, not confirmed product priorities, shipped capabilities, or how resources are ultimately allocated.
The posting set combines product, engineering, sales, and internal operations roles, so some signals may reflect organizational scaling as well as product investment.
Descriptions cite large usage and infrastructure figures, but postings do not establish how those figures are measured or how they compare over time.