Perplexity

Perplexity is AI for people who expect more. This role brings that same standard to how our data team works, with AI at the center of everything we do.We're looking for someone who's been a great data scientist, analytics engineer, or data engineer: the kind of person who knows which metric actually matters, can design an A/B test that answers the real question, has gone deep on a data model because something didn't add up, and has decided that the highest-leverage thing they can do next is build AI systems that fundamentally change how data science gets done.You'll build AI agents and internal systems that can increasingly handle end-to-end analysis workflows: forming hypotheses, writing and running queries, interpreting results, and drafting recommendations with the right evaluation, review, and guardrails. You'll build the retrieval infrastructure and evaluation loops that let AI systems query the warehouse reliably. You'll create workflows that detect, diagnose, and help fix data issues before they become company-wide problems. You'll build the infrastructure that multiplies what a small data team can ship.You'll join a data team that's already using AI across its work. The next step is turning those individual workflows into scalable systems, shared tools, and an AI-native operating model that becomes a benchmark for how modern data teams work.<h2>What You'll Do</h2><ul><li>Build AI agents that do data science - not just SQL copilots, but systems that can safely explore data, form hypotheses, run queries, interpret results, and generate actionable recommendations with clear evaluation and human review loops.</li><li>Make AI systems query the warehouse reliably - build the retrieval infrastructure and evaluation loops that let agents use our semantic context and metadata accurately.</li><li>Accelerate the AI-native data workflow - turn the best existing AI-assisted workflows into repeatable systems, reusable tools, and patterns the whole data team can adopt.</li><li>Automate the data lifecycle - build self-healing pipelines, automated dbt model generation and validation, data quality agents, and diagnosis workflows that reduce manual firefighting.</li><li>Ship AI-powered experiment analysis - build agents that interpret A/B test results, flag statistical issues, identify likely drivers, and draft ship/no-ship recommendations.</li><li>Turn the data team into a product team - build internal data products that stakeholders use every day, replacing ad hoc requests with self-serve AI interfaces.</li><li>Own the full lifecycle - identify high-leverage problems, prototype with LLMs, evaluate accuracy, design the UX, ship to production, and monitor quality over time.</li></ul><h2>What We're Looking For</h2><ul><li>6+ years in data science, analytics engineering, data engineering, or a related role. You've been close enough to real data work to know what should and should not be automated.</li><li>Deep SQL and analytics judgment - you can reason through metrics, experiments, data models, and messy warehouse reality without relying on a tool to think for you.</li><li>Strong product sense - you understand what stakeholders actually need, what makes a workflow adoptable, and how to turn a prototype into a product people use.</li><li>Production-oriented Python ability - you can build and ship working tools, wrangle APIs, evaluate model outputs, deploy services, and write code others can maintain.</li><li>Hands-on LLM experience - you've built with frontier models, agents, RAG systems, evals, or AI-powered workflows and have opinions about where they work and where they fail.</li><li>Pipeline and modeling fluency - you've worked with dbt, warehouse schemas, data quality issues, and the practical tradeoffs behind durable data systems.</li><li>Builder mentality - you see a manual process and immediately think about how to systematize it. You ship fast, measure quality, and iterate.</li><li>Autonomy - this is a new function. You'll help define the roadmap as much as execute it.</li></ul><h2>Bonus</h2><ul><li>Experience building production AI agents or agent evaluation systems.</li><li>Experience with Snowflake, semantic layers, or metadata systems.</li><li>Experience building internal tools, Slack bots, CLIs, or developer productivity products that people actually used.</li><li>Strong experimentation background, including metric design and statistical interpretation.</li><li>Experience with BI tools and the judgment to know what should be automated versus kept human-reviewed.</li><li>Early-stage startup experience.</li></ul><h2>Why This Role</h2><ul><li>Set the standard for the industry - Perplexity's data team is already using AI across its work. You'll turn that into something other data teams look to as the benchmark.</li><li>Build AI with AI - Perplexity builds AI for people who expect more. You'll bring that same level of ambition to how the company works with data.</li><li>Frontier models, day one - you're at an AI company with access to frontier infrastructure and people who deeply understand what's possible.</li><li>Massive leverage - the systems you build will multiply the output of every data team member and every stakeholder who needs data.</li><li>Direct impact - small team, no layers of approval. Idea to shipped system in days, not quarters.</li></ul>

Member of Data Staff (AI Builder)

<h2>About the Role</h2>The Data Platform team owns the end-to-end data lifecycle at Perplexity, from ingestion through processing, storage, and serving, powering product features, analytics, experimentation, AI workloads, and the company’s data lake.The team defines the architecture for batch and streaming systems, the orchestration and observability stack, and a self-serve data platform, while thoughtfully combining platforms such as Databricks and Snowflake with open-source technologies including Spark, Kafka, Flink, Airflow, Dagster, dbt, Iceberg, Delta Lake, and ClickHouse.In this senior/staff role, you will shape architecture, set standards, and drive the long-term technical direction of Perplexity’s data ecosystem.<h2>Key Responsibilities</h2><ul><li>Design and operate large-scale batch and streaming data pipelines that directly power Perplexity product features, AI training and evaluation workflows, analytics, and experimentation.</li><li>Build event-driven and streaming systems (Kafka, Kinesis, PubSub, or similar) for real-time ingestion, transformation, and delivery, alongside batch frameworks for backfills, aggregations, and offline computation.</li><li>Lead the architecture of data orchestration using tools like Airflow or Dagster, owning scheduling, dependency management, retries, SLAs, and end-to-end observability for critical data flows.</li><li>Set and enforce guarantees for data correctness, freshness, lineage, and recoverability, designing systems that handle rapid scale growth, partial failures, and evolving schemas without disrupting AI workloads or product experiences.</li><li>Build self-serve data platforms that let engineers, data scientists, and analysts safely discover data, define contracts, and create and operate their own pipelines with minimal friction.</li><li>Improve developer experience through better abstractions, opinionated paved paths, and standards for data modeling, testing, validation, and deployment, treating the data platform as a product used by many teams.</li><li>Drive architectural decisions across storage, compute, orchestration, and data APIs, partnering closely with product engineering and data science to align the data ecosystem with Perplexity’s roadmap.</li><li>Mentor engineers, review designs, and raise the technical bar for data infrastructure through thoughtful feedback, documentation, and hands-on collaboration.</li></ul><h2>Qualifications</h2><ul><li>5+ years (Senior) or 8+ years (Staff) of software engineering experience.</li><li>Strong experience building production data infrastructure systems.</li><li>Hands-on experience with batch and/or streaming data processing at scale.</li><li>Deep familiarity with data orchestration systems (Airflow, Dagster, or similar).</li><li>Proficiency in Python and at least one additional backend language (Go, TypeScript, etc.).</li><li>Strong systems thinking around reliability, latency, cost, and complexity tradeoffs.</li><li>Experience supporting ML/AI workflows, training pipelines, or evaluation systems.</li><li>Familiarity with data quality, lineage, observability, and governance tooling.</li><li>Prior ownership of internal platforms used by many teams.</li></ul>If you’re excited about this role, we encourage you to apply even if your experience doesn’t match every qualification listed above.

Member of Technical Staff (Software Engineer, Data Platform)

We are looking for an AI Infra engineer to join our growing team. We work with Kubernetes, Slurm, Python, C++, PyTorch, and primarily on AWS. As an AI Infrastructure Engineer, you will be partnering closely with our Inference and Research teams to build, deploy, and optimize our large-scale AI training and inference clusters. <h2>Responsibilities</h2><ul><li>Design, deploy, and maintain scalable Kubernetes clusters for AI model inference and training workloads</li><li>Manage and optimize Slurm-based HPC environments for distributed training of large language models</li><li>Develop robust APIs and orchestration systems for both training pipelines and inference services</li><li>Implement resource scheduling and job management systems across heterogeneous compute environments</li><li>Benchmark system performance, diagnose bottlenecks, and implement improvements across both training and inference infrastructure</li><li>Build monitoring, alerting, and observability solutions tailored to ML workloads running on Kubernetes and Slurm</li><li>Respond swiftly to system outages and collaborate across teams to maintain high uptime for critical training runs and inference services</li><li>Optimize cluster utilization and implement autoscaling strategies for dynamic workload demands</li></ul><h2>Qualifications</h2><ul><li>Strong expertise in Kubernetes administration, including custom resource definitions, operators, and cluster management</li><li>Hands-on experience with Slurm workload management, including job scheduling, resource allocation, and cluster optimization</li><li>Experience with deploying and managing distributed training systems at scale</li><li>Deep understanding of container orchestration and distributed systems architecture</li><li>High level familiarity with LLM architecture and training processes (Multi-Head Attention, Multi/Grouped-Query, distributed training strategies)</li><li>Experience managing GPU clusters and optimizing compute resource utilization</li></ul><h2>Required Skills</h2><ul><li>Expert-level Kubernetes administration and YAML configuration management</li><li>Proficiency with Slurm job scheduling, resource management, and cluster configuration</li><li>Python and C++ programming with focus on systems and infrastructure automation</li><li>Hands-on experience with ML frameworks such as PyTorch in distributed training contexts</li><li>Strong understanding of networking, storage, and compute resource management for ML workloads</li><li>Experience developing APIs and managing distributed systems for both batch and real-time workloads</li><li>Solid debugging and monitoring skills with expertise in observability tools for containerized environments</li></ul><h2>Preferred Skills</h2><ul><li>Experience with Kubernetes operators and custom controllers for ML workloads</li><li>Advanced Slurm administration including multi-cluster federation and advanced scheduling policies</li><li>Familiarity with GPU cluster management and CUDA optimization</li><li>Experience with other ML frameworks like TensorFlow or distributed training libraries</li><li>Background in HPC environments, parallel computing, and high-performance networking</li><li>Knowledge of infrastructure as code (Terraform, Ansible) and GitOps practices</li><li>Experience with container registries, image optimization, and multi-stage builds for ML workloads</li></ul><h2>Required Experience</h2><ul><li>Demonstrated experience managing large-scale Kubernetes deployments in production environments</li><li>Proven track record with Slurm cluster administration and HPC workload management</li><li>Previous roles in SRE, DevOps, or Platform Engineering with focus on ML infrastructure</li><li>Experience supporting both long-running training jobs and high-availability inference services</li><li>Ideally, 3-5 years of relevant experience in ML systems deployment with specific focus on cluster orchestration and resource management</li></ul>

Member of Technical Staff (AI Infrastructure Engineer)

We are looking for an AI Infra engineer to join our growing team. We work with Kubernetes, Slurm, Python, C++, PyTorch, and primarily on AWS. As an AI Infrastructure Engineer, you will be partnering closely with our Inference and Research teams to build, deploy, and optimize our large-scale AI training and inference clusters<h2>Responsibilities</h2><ul><li>Design, deploy, and maintain scalable Kubernetes clusters for AI model inference and training workloads</li><li>Manage and optimize Slurm-based HPC environments for distributed training of large language models</li><li>Develop robust APIs and orchestration systems for both training pipelines and inference services</li><li>Implement resource scheduling and job management systems across heterogeneous compute environments</li><li>Benchmark system performance, diagnose bottlenecks, and implement improvements across both training and inference infrastructure</li><li>Build monitoring, alerting, and observability solutions tailored to ML workloads running on Kubernetes and Slurm</li><li>Respond swiftly to system outages and collaborate across teams to maintain high uptime for critical training runs and inference services</li><li>Optimize cluster utilization and implement autoscaling strategies for dynamic workload demands</li></ul><h2>Qualifications</h2><ul><li>Strong expertise in Kubernetes administration, including custom resource definitions, operators, and cluster management</li><li>Hands-on experience with Slurm workload management, including job scheduling, resource allocation, and cluster optimization</li><li>Experience with deploying and managing distributed training systems at scale</li><li>Deep understanding of container orchestration and distributed systems architecture</li><li>High level familiarity with LLM architecture and training processes (Multi-Head Attention, Multi/Grouped-Query, distributed training strategies)</li><li>Experience managing GPU clusters and optimizing compute resource utilization</li></ul><h2>Required Skills</h2><ul><li>Expert-level Kubernetes administration and YAML configuration management</li><li>Proficiency with Slurm job scheduling, resource management, and cluster configuration</li><li>Python and C++ programming with focus on systems and infrastructure automation</li><li>Hands-on experience with ML frameworks such as PyTorch in distributed training contexts</li><li>Strong understanding of networking, storage, and compute resource management for ML workloads</li><li>Experience developing APIs and managing distributed systems for both batch and real-time workloads</li><li>Solid debugging and monitoring skills with expertise in observability tools for containerized environments</li></ul><h2>Preferred Skills</h2><ul><li>Experience with Kubernetes operators and custom controllers for ML workloads</li><li>Advanced Slurm administration including multi-cluster federation and advanced scheduling policies</li><li>Familiarity with GPU cluster management and CUDA optimization</li><li>Experience with other ML frameworks like TensorFlow or distributed training libraries</li><li>Background in HPC environments, parallel computing, and high-performance networking</li><li>Knowledge of infrastructure as code (Terraform, Ansible) and GitOps practices</li><li>Experience with container registries, image optimization, and multi-stage builds for ML workloads</li></ul><h2>Required Experience</h2><ul><li>Demonstrated experience managing large-scale Kubernetes deployments in production environments</li><li>Proven track record with Slurm cluster administration and HPC workload management</li><li>Previous roles in SRE, DevOps, or Platform Engineering with focus on ML infrastructure</li><li>Experience supporting both long-running training jobs and high-availability inference services</li><li>Ideally, 3-5 years of relevant experience in ML systems deployment with specific focus on cluster orchestration and resource management</li></ul>

We build and run the inference engine behind every Perplexity query and deploy dozens of model architectures at scale with tight latency and cost budgets. Our stack is Rust, Python, CUDA, and CuTe DSL - and we need another engineer to join us.<h2>What you will work on</h2>Examples of real work the team does:<ul><li>New models support. Support transformer-based retrieval, text-generation, and multimodal models in our inference infrastructure, from weight loading, request scheduling and KV-cache management to support in API Gateway.</li><li>GPU kernels migration to CuTe DSL. Port our in-house CUDA kernels to NVIDIA's CuTe DSL so they run on GB200 today and are portable to Vera Rubin racks tomorrow.</li><li>Rust-native serving runtime. Develop our internal Rust-based inference server to solve all Python pains and keep up with rapidly growing traffic.</li><li>Performance optimisation. Profile and fix bottlenecks from network ingress through continuous batching and GPU kernel interleaving.</li><li>Reliability and observability. Build dashboards, alerts, and automated remediation so we catch regressions before users do. Respond to and learn from production incidents.</li></ul><h2>Who we're looking for</h2><ul><li>Deep experience with GPU programming and performance work (CUDA, Triton, CUTLASS, or similar). Any other deep systems programming experience is a plus.</li><li>You understand modern LLM architectures and are able to bring them up reliably in a production environment.</li><li>You've built and operated production distributed systems under real load - ideally performance-critical ones.</li><li>Comfortable working across languages and layers: Rust for the serving runtime, Python for model code, CUDA/CuteDSL for kernels.</li><li>You own problems end-to-end. You can read a research paper on Monday, write a kernel on Wednesday, and debug a production incident on Friday.</li><li>Self-directed. You do well in fast-moving environments where the path forward isn't laid out for you.</li></ul><h2>Good if you touched any of</h2><ul><li>ML compilers and framework internals: PyTorch internals, torch.compile, custom operators.</li><li>Distributed GPU communication: NCCL, NVLink, InfiniBand, RDMA libraries, model/tensor parallelism.</li><li>Low-precision inference: INT8/FP8/FP4 quantization, mixed-precision serving.</li><li>Profiling and debugging tools: Nsight Compute/Systems, CUDA-GDB, PTX/SASS analysis.</li><li>Container orchestration: Kubernetes, GPU scheduling, autoscaling inference workloads.</li></ul><h2>Qualifications</h2><ul><li>3+ years of professional software engineering experience with meaningful work on ML inference or high-performance systems.</li><li>Familiarity with at least one deep learning framework (PyTorch, JAX, TensorFlow).</li><li>Understanding of GPU architectures (memory hierarchy, warp scheduling, tensor cores).</li><li>Understanding of common LLM architectures and inference optimization techniques (e.g. quantization, speculative decoding, prefill-decode disaggregation).</li></ul>

Perplexity

Open jobs at Perplexity

Member of Data Staff (AI Builder)

Member of Technical Staff (Software Engineer, Data Platform)

Member of Technical Staff (AI Infrastructure Engineer)

Member of Technical Staff (AI Infrastructure Engineer)

Member of Technical Staff (AI Inference Engineer)

Working for Perplexity?

Featured companies