Staff Software Engineer, Data
Stability AI, Remote
The team
The Data team builds and runs the systems that make Stability AI's data usable, both to operate the business and to train our models. We're a small, collaborative team that works closely with Finance, Marketing, SRE, Security, Product and Research. We focus on outcomes rather than hours, and you'll have deep technical access and the freedom to build.
Our stack: AWS (S3, Athena, Glue, EMR, Step Functions, EventBridge), Iceberg, dbt, Postgres, Terraform, Helm, GitHub Actions, QuickSight, Stripe, Auth0, HubSpot, Customer.io, PostHog, OpenTelemetry and SLURM — or similar. We care about the judgement more than the logo.
The Work
You'll own our company data platform end to end: the AWS data lake, the transformation layer, the integrations between our business systems, and the Kubernetes-based infrastructure they run on. The data that runs the business (billing, usage telemetry, identity, CRM, ERP, cloud cost) lives in a dozen systems that were never wired together. You'll be the engineer who connects them and keeps them reliable.
You'll also be the data engineer for our research teams across audio, video, image and 3D, building the pipelines that turn raw media into training-ready datasets. This is a hands-on role with real ownership, not a data-prep or analyst seat.
- Own the company data platform end to end: the S3/Athena/Glue lake, the dbt project and its daily build, ingestion from Stripe, Auth0, HubSpot, OpenTelemetry and product events, and the Postgres serving layer for dashboards and AI agents
- Build and operate integrations between billing, CRM, ERP, identity and telemetry systems so revenue, usage, cost and customer identity reconcile automatically instead of through spreadsheets and tickets
- Design and maintain the warehouse layer: dimensional models, data contracts with product engineering for event schemas, quality tests, lineage and a catalog
- Enforce PII tiering (open vs. restricted schemas) through access grants, working with Security and Legal
- Process data at scale: billions of telemetry rows, terabyte log streams, Iceberg tables, partitioning and cost control
- Own orchestration and infrastructure: orchestration on EKS, Helm, Terraform, CI/CD, monitoring and alerting, and join the business-hours on-call rotation for the daily pipelines
- Deliver operational reporting for Finance and leadership: revenue by business line, credits and prepaid liability, and cloud cost by department
- Support the research teams across every modality: source, license-check, ingest and version large training datasets for audio, video, image and 3D