Software Engineer, Knowledge Engine – Query Engine

Exa is an applied AI lab building a search engine unlike the world has ever seen. We build massive-scale infra to crawl the entire web, train state-of-the-art embedding models to process it, and design super high performant vector databases to retrieve over it. We now power search for Cursor, Cognition, HubSpot, and over 400,000 developers and have raised $350m from Lightspeed, Benchmark, and a16z.

Our ultimate goal is to build perfect search over all the world's information, far beyond Google. If you want to build massive-scale ML systems that will define the way the new AI world consumes information, this is the place for you.

The Knowledge Platform team builds Exa's knowledge engine. We continuously synthesize information from across the web and beyond into one of the world's largest open-domain knowledge graphs, with hundreds of billions of evidence-backed claims. The work has two halves: construction pipelines that decide what to trust, and retrieval engines that decide how best to search it. Our goal is to give agents answers to any factual question about the world with unprecedented precision and completeness.

In this role, you'll own the query engine that searches hundreds of billions of claims across a variety of retrieval backends, from a distributed graph database to vector, full-text, time-series and geospatial indexes.

This is greenfield work with a massive scope. No off-the-shelf system answers precise questions over hundreds of billions of evidence-backed claims at agent speed, so you'll make foundational decisions, from storage layout and indexing across engines to the query language and the planner that routes work between them, that every query depends on.

What you'll do

• Own storage and indexing across backends: data layout, partitioning and indexes for graph, vector, full-text, time-series and geospatial retrieval, with bulk snapshot builds and live ingest that never take the system down.
• Decide how best to search: cost-based planning from cardinality estimates that chooses, for each question, which backends to use, in what order and how to join their results, executed as traversals, retrievals, filters and aggregations in one plan.
• Make it fast at scale: serve thousands of queries per second, profiling and cutting p50/p99 latency, memory and cost per query under high concurrent load, from on-disk layout to network fan-out.
• Keep answers consistent while the data converges: the graph is fed by live updates and bulk snapshots at once and is eventually consistent by design, so you'll own what a query may see while it converges: consistent snapshot promotion, ordered application of live updates and stable pagination across versions.
• Measure precision and completeness: end-to-end checks every deployment must pass, plus evals and metrics that show when a plan returns wrong results or misses ones it should have found.

What we're looking for

• You've worked deep inside a distributed database, query engine, search engine or storage system.
• Strong grasp of indexing, query planning and optimization, and storage formats.
• Solid grasp of distributed systems fundamentals: consensus, replication, sharding and fault tolerance.
• Strong in a systems language (Go, Rust or C++).
• You profile before you optimize, and you've shipped measured latency wins in production.
• You obsess over correctness, reliability and data quality as much as scale.
• Bonus
• Graph databases, or combining vector, full-text, time-series and geospatial retrieval
• Compilers, query language design or cost-based optimizers
• Example projects
• Build a system that retrieves and reasons over everything known about any company or person in the world, answering any factual question precisely and completely.
• Answer precise multi-hop questions over hundreds of billions of claims at thousands of QPS with interactive latency.
• Build a planner that picks the right mix of graph, vector, full-text and geospatial retrieval for any question, and proves it chose well.
• Make new knowledge queryable seconds after it's synthesized, without ever serving an inconsistent result.

Logistics

Location:

This is an in-person opportunity in San Francisco.

Visas:

We're happy to sponsor international candidates (e.g., STEM OPT, OPT, H1B, O1, E3). While we cannot guarantee your visa, we have historically been successful in sponsoring candidates from all over the world. If you receive an offer, our team will work hard to get you a visa.