SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Eventual is building the data infrastructure layer for Physical AI and robotics. The company has raised $30M from top-tier investors (Felicis, CRV, Y Combinator) and counts FAANG companies, Mobileye, and TogetherAI as customers. The core product is Daft, an open-source engine purpose-built for multimodal AI data at petabyte scale, handling video, lidar, radar, and sensor data that traditional platforms like Databricks and Snowflake cannot efficiently process.
As a Software Engineer on the Data Systems team, you will own critical infrastructure components that power real-time indexing, distributed storage, querying, and GPU dataloading for frontier robotics and Physical AI labs. The role spans the full stack: compute infrastructure, data storage/querying layers, and model training/deployment pipelines.
Key responsibilities include: (1) Multimodal Storage—designing and optimizing data lake architectures using Apache Parquet, Apache Iceberg, and similar columnar formats for high-dimensional video, lidar, and sensor logs; (2) Query Engine—building powerful multi-stage query systems with database fundamentals like partitioning, indexing, query planning, vector search, and LLM/VLM-based perception predicates; (3) Dataloading—improving memory stability, throughput, and zero-copy data flow through streaming computation and line-rate CUDA tensor delivery to GPUs.
You will work in a small, experienced engineering team that values technical autonomy, deep execution, and solving hard distributed systems problems. The team is based in the SF Mission District with a 4-day/week in-office expectation.
Required: 3+ years building resilient, high-throughput distributed systems or database engines in Rust or C++. Deep experience with engine internals (vectorized execution, query planning/optimization, distributed task scheduling, zero-copy networking). Practical exposure to cloud infrastructure scaling (AWS S3) and heavy-compute data pipelines. Bonus: CUDA, GPU streaming, video decoding frameworks. You should thrive in an autonomous, fast-paced startup environment.
