SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Anthropic's RL Data Platform team builds the infrastructure that powers human feedback collection and data pipeline systems for training Claude. This is a full-stack, ownership-heavy role on a small, senior engineering team.
You'll design and ship web interfaces used by thousands of expert annotators, build backend services and data pipelines that support them, and work directly with RL researchers to understand their data needs. The role spans the full stack: TypeScript/React frontends, Python backends, and data infrastructure. You'll scope your own projects, make architectural decisions, and see them through to production.
Key responsibilities include:
- Design, build, and operate feedback and data collection interfaces for human annotators, domain experts, and researchers
- Build and maintain backend services, APIs, and pipelines that route model samples to humans and return structured feedback to training
- Own reliability, latency, and usability of systems running continuously against live model endpoints
- Partner with RL researchers to translate data needs into well-scoped collection campaigns and tooling
- Build dashboards, monitoring, and inspection tools so researchers can self-serve data quality and throughput visibility
- Identify and remove bottlenecks between "we want this data" and "it's in the training mix"
Required qualifications: strong full-stack engineering with production TypeScript/React and Python experience; experience designing and operating backend services and data pipelines; track record of owning projects end-to-end from ambiguous brief to production; comfort working with technical stakeholders whose needs change frequently; effective use of AI tools; care about societal impacts of your work.
Preferred: experience with annotation/labeling/evaluation tooling, RLHF or human-feedback pipelines, shipping researcher-facing internal tools, running experiments on data collection interfaces, working with crowdworker platforms at scale, familiarity with LLM training and evaluation.
