Exa is building a search engine from scratch to serve every AI application. We build massive-scale infrastructure to crawl the web, train state-of-the-art embedding models to index it, and develop super high performant vector databases in Rust to search over it. We also own a $5m H200 GPU cluster and routinely run batchjobs with tens of thousands of machines. This isn't your average startup :)
As a Data Engineer, you'll architect and build the data infrastructure that powers everything we do—from crawling billions of pages to training our embedding models to serving real-time search. You'll have enormous autonomy in designing systems that scale to hundreds of petabytes. If you've ever wanted to build data pipelines at a scale that most companies only dream about, this is your chance.
Desired Experience
Deep understanding of lakehouse architectures (Delta Lake, Iceberg, Hudi) and when to use them
Experience building and operating large-scale distributed data processing pipelines
Hands-on experience with streaming data systems (Kafka, Flink, or similar)
Familiarity with Ray, Spark, or ClickHouse at production scale
An obsessive focus on reliability and building systems that don't page you at 3am
Bonus Points
Experience with Lance or other vector-native storage formats
Background in GPU-accelerated data processing (RAPIDS, cuDF)
Example Projects
Design a lakehouse architecture that handles 100+ PB of web crawl data
Build streaming pipelines that process billions of documents per day for real-time indexing
Architect the data layer for our embedding training infrastructure on Ray
Scale our ClickHouse deployment to handle analytical queries across petabytes of search logs
This is an in-person opportunity in San Francisco. We're happy to sponsor international candidates (e.g., STEM OPT, OPT, H1B, O1, E3).
Top Skills
What We Do
Exa was built with a simple goal — to organize all knowledge. After several years of heads-down research, we developed novel representation learning techniques and crawling infrastructure so that LLMs can intelligently find relevant information.
