Software Engineer, Knowledge Engine – Query Engine

Exa•San Francisco, CA

1d•$180,000 - $350,000•Onsite

About The Position

Exa is hiring a Software Engineer for its Knowledge Platform team to own the query engine for a knowledge graph containing hundreds of billions of evidence-backed claims. The engine searches claims across retrieval backends including distributed graph, vector, full-text, time-series, and geospatial systems. This is greenfield work involving foundational decisions about storage, indexing, query language, and planning so agents can answer factual questions with precision and completeness. The role includes designing storage layouts and indexes, building query plans that select and combine retrieval backends, and improving performance under high concurrent load. The engineer will also define how queries observe data while live updates and bulk snapshots converge, and develop deployment checks, evaluations, and metrics for precision and completeness. The opportunity is in person in San Francisco. Exa is willing to sponsor international candidates, though visa sponsorship is not guaranteed.

Requirements

Deep experience working inside a distributed database, query engine, search engine, or storage system.
Strong understanding of indexing, query planning and optimization, and storage formats.
Solid understanding of distributed systems fundamentals, including consensus, replication, sharding, and fault tolerance.
Strong proficiency in a systems language such as Go, Rust, or C++.
A profile-first approach to optimization and a record of shipping measured latency improvements in production.
Strong focus on correctness, reliability, and data quality as well as scale.

Nice To Haves

Experience with graph databases or combining vector, full-text, time-series, and geospatial retrieval.
Experience with compilers, query language design, or cost-based optimizers.

Responsibilities

Own storage and indexing across graph, vector, full-text, time-series, and geospatial backends, including data layout, partitioning, indexes, bulk snapshot builds, and live ingest.
Build cost-based query planning using cardinality estimates to select retrieval backends, determine their order, and join results through a unified plan.
Improve performance at scale by profiling and reducing p50/p99 latency, memory use, and cost per query under high concurrent load.
Define query consistency during convergence of live updates and bulk snapshots, including snapshot promotion, ordered update application, and stable pagination across versions.
Create end-to-end deployment checks, evaluations, and metrics to identify incorrect or incomplete query results.