职位描述
Help us build the multimodal data stack for physical AI
At Rerun, we’re building a new multimodal data stack for physical AI. Large parts of the physical-world economy will be transformed by these technologies, and we're building the data infrastructure and tools that enable it to happen. Data volumes grow big quickly with these applications and performance matters everywhere; from micro interactions to high throughput jobs.
With multimodal data stack we mean tools for extraction, ingestion, storage, querying, streaming, and visualization of temporal multimodal data. Rerun’s open source project is already loved by some of the best teams in the world, and focuses on visualizing and handling multimodal logs. You'll join a newly formed team building out Rerun's commercial offering for ingesting, indexing and querying multimodal data at scale.
What you'll work on
- You will work on a cloud-native analytical database system built for multimodal (often time-series) data.
- The system is built in Rust and makes extensive use of Apache Arrow. It supports specialised queries to handle robotics-style multimodal logs (very sparse and wide tables) and domain specific operators (e.g. related to 3D transforms).
- Note: prior robotics knowledge is not necessary
- You will design, implement and operate the core pieces that make sure this data-intensive system is highly scalable, reliable and observable.