Member of Technical Staff (Applied AI Engineer, Agent Capabilities)
PerplexitySan Francisco, California, USposted 20d ago
At a glance
Build and ship frontier agent capabilities at Perplexity, turning emerging model behaviors into reliable, scalable products. Own the full lifecycle from evaluation and prototyping through production monitoring and iteration.
Summarized by AI from the original posting
What you'll do
- Evaluate frontier models against real user tasks and identify useful behaviors and failure modes
- Turn model advances into reliable, scalable production agent systems
- Improve agent planning, tool use, context management, error recovery, and long-running task execution
- Design scalable agent capabilities using state-of-the-art ML and LLM techniques
- Own agent behavior and capabilities from user-facing products to backend services
- Define offline and online evaluations for task completion, correctness, safety, latency, cost, and user satisfaction
- Build secure, observable, and reliable agent systems with permissions and safeguards
- Develop tracing, replay, and monitoring infrastructure
- Apply advances in models, inference, evaluation, and agent architecture to production
- Set technical direction, mentor others, and provide technical leadership
Requirements
- 01Track record of building and owning robust AI-powered, large-scale, user-facing or data-intensive products
- 02Strong software engineering fundamentals
- 03Experience building and operating AI/ML products, backend services, or distributed systems at scale
- 04Experience owning the AI product lifecycle
- 05Experience with agent harnesses, tool use, context engineering, model evaluation, browser automation, or long-running task execution
- 06Strong product judgment and execution
- 07Genuine interest in frontier AI capabilities and agent systems
- 08Experience with LLM context engineering or harness engineering
- 09Experience with subagents, coding assistants, or autonomous task execution
- 10Experience building agent permissions, safeguards, evaluation infrastructure, or production observability systems
- 11Experience with mid-training, post-training, or reinforcement learning
- 12AI/ML research experience demonstrated through publications, open-source contributions, or other meaningful research impact
Skills
Full description
Perplexity Computer is one of the defining products of the new era of agentic AI. Millions of people use Perplexity to transform knowledge into action, and the Agent Capabilities team sits at the intersection of frontier AI research and product innovation, building the foundations that shape how users and agents solve increasingly complex tasks.
As every major breakthrough in AI models creates new possibilities, the Agent Capabilities team is responsible for turning frontier AI breakthroughs into reusable product capabilities. We are often the first to evaluate emerging model capabilities, determine where they create real user value, and transform them into reliable, scalable, high quality experiences for both users and agents. This is a highly leveraged role with broad ownership at the intersection of frontier AI research, agent systems, platform engineering, and product innovation.
Tech Stack: Python | Go | Rust | PostgreSQL | DynamoDB | AWS | TypeScript
Why Perplexity is different
Craftsmanship. We build high quality, tasteful products targeting both the AI native and AI curious.
Ownership. You identify the problem, design the solution and ship it.
Entrepreneurship. We think like founders, act with urgency, and hustle to deliver for each other and our users.
Scholarship. Work among highly talented peers, pursuing knowledge and truth, upleveling ourselves, our teams, and our products.
Partnership. We amplify each others' strengths, break down silos, and give selflessly to help our colleagues deliver excellence.
What you'll do
Evaluate frontier models against real user tasks, identify useful behaviors and failure modes, and turn the most promising advances into production agent systems. Own the lifecycle from rapid prototyping and evaluation through launch, monitoring, and iteration.
Improve agents’ ability to plan, use tools, manage context, recover from errors, and complete long-running tasks reliably.
Apply state of the art ML and LLM techniques to design scalable agent capabilities such as skills, plugins, artifact generation, tools integrate and use, auto-research, and multi-agent collaboration. Shape the architecture, abstractions, and product experiences that enable both users and agents to compose increasingly sophisticated solutions for real-world tasks.
Own agent behavior and capabilities end-to-end, from user-facing products and interfaces to backend services. Define offline and online evaluations for task completion, correctness, safety, latency, cost, and user satisfaction. Iteratively improve across models, prompts, harnesses, and products for different problem spaces.
Build secure, observable, and reliable agent systems, including permissions and safeguards for sensitive actions. Develop tracing, replay, and monitoring infrastructure that makes agent failures reproducible and actionable.
Collaborate closely with PM, Data Science, Research, to identify high-impact opportunities in understanding and validating emerging model capabilities, and turn complex agent behaviors into simple, reliable product experiences.
Apply relevant advances in models, inference, evaluation, and agent architecture when they produce measurable improvements in production performance. Set technical direction on ambiguous problems and raise the bar through design reviews, mentorship, and technical leadership.
Qualifications
Typically 6+ years of professional software engineering experience, with a track record of building and owning robust AI-powered, large-scale, user-facing or data-intensive products. Exceptional candidates with less experience and an outstanding record of impact are encouraged to apply.
Strong software engineering fundamentals, with experience building and operating AI/ML products, backend services, or distributed systems at scale.
Experience owning the AI product lifecycle, including data analysis, rigorous evaluation, production monitoring, and iterative improvement. Able to define metrics and use production data and user feedback to guide decisions.
Practical experience in one or more relevant areas, such as agent harnesses, tool use, context engineering, model evaluation, browser automation, or long-running task execution.
Strong product judgment and execution: you can translate ambiguous user needs into applied AI or ML problems and ship durable solutions with measurable user impact.
Genuine interest in frontier AI capabilities, agent systems, and excitement for rapidly exploring, evaluating, and productizing new model behaviors.
Nice to have
Experience with LLM context engineering or harness engineering, experience with subagents, coding assistants, long-running or autonomous task execution.
Deep familiarity with the strengths and limitations of current model families across reasoning, tool use, context management, and long-horizon tasks.
Experience building agent permissions, safeguards, evaluation infrastructure, or production observability systems.
Experience with mid-training, post-training, or reinforcement learning for frontier or open-source models, along with a strong understanding of model strengths and limitations across reasoning, tool use, context management, and long-horizon tasks.
AI/ML research experience demonstrated through publications, open-source contributions, or other meaningful research impact.
Time spent at a fast-growing startup or on a high-ownership engineering team.