Perplexity Tech Stack

AI-powered answer engine with real-time source verification

Software Development San Francisco, California 201–500 employees Founded 2022 Privately Held

Perplexity operates a search and answer platform built on a heavy machine-learning stack: PyTorch, Triton, CUDA, TensorFlow, ONNX for model inference, plus Next.js and React on the frontend. The company is actively adopting RAG (retrieval-augmented generation), Kafka, and infrastructure-as-code (Terraform, dbt, Snowflake, Databricks), while simultaneously scaling inference bottlenecks and building agentic products—a set of challenges that explains their senior-heavy engineering hiring and focus on LLM deployment and CI/CD pipeline maturity.

Tech Stack 108 technologies

Core StackPython Next.js TypeScript Docker AWS Kubernetes PostgreSQL Rust C++ PyTorch TensorFlow Tailwind CSS React DynamoDB Redis FastAPI Terraform Ansible dbt Snowflake Databricks Triton CUDA ONNX CSS Chromium Chrome WebView Slurm SQL+75 more

AdoptingTerraform SonarQube Kafka RAG dbt Kinesis Google Cloud Pub/Sub

What Perplexity Is Building

◆Challenges

Improving search quality
Bottlenecks in inference stack
Diagnosing performance bottlenecks
Building core infrastructure for agentic products
Scalable llm deployment
Data security compliance risks
Improving ci/cd pipelines
Scaling data pipelines for millions of users
Reducing manual work in incident response
Scaling data team capacity

▲Active Projects

Core search components
Develop apis for ai inference for internal and external customers
Scalable legal infrastructure
Endpoint management
Rag pipelines for grounding and answer generation
Comet browser
Ci/cd pipeline implementation
Develop llm-as-a-judge systems
Ai-readable data warehouse
Unified file layer

Hiring Activity

Decelerating100 roles · 25 in 30d

Department

Engineering

Data

Security

Finance

Support

Legal

Marketing

Ops

Seniority

Senior

Mid

Lead

Manager

Intern

Junior

Staff

Notable leadership hires: Site Lead

Company intelligence

Find more companies like Perplexity by tech stack, pain points and active projects

Get started free

About Perplexity

Perplexity is an answer engine that combines real-time web search with AI-powered reasoning to deliver sourced responses. Founded in 2022 and based in San Francisco, the company operates a 201–500-person organization with a backend and infrastructure focus: 53 engineers, 8 data specialists, and 7 security staff actively hiring. The platform relies on Python, Kubernetes, and a complex machine-learning inference stack, with active development on core search components, RAG pipelines for grounding answers, LLM-as-a-judge systems for quality control, and a companion browser product. The company also offers APIs for AI inference to external customers, signaling a move toward a platform business model alongside the direct consumer product.

HeadquartersSan Francisco, California

Company Size201–500 employees

Founded2022

Hiring MarketsUnited States, Germany, United Kingdom, Serbia, Japan

Frequently Asked Questions

What tech stack does Perplexity use?

Python, Next.js, TypeScript, PyTorch, Triton, CUDA, TensorFlow, Kubernetes, PostgreSQL, Redis, AWS, DynamoDB, Snowflake, and Databricks. Currently adopting Kafka, Terraform, dbt, and RAG frameworks.

What is Perplexity working on?

Core search infrastructure, RAG pipelines for answer grounding, LLM-as-a-judge systems for quality control, a companion browser, APIs for external AI inference, and scalable legal infrastructure to support platform growth.