SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Rhoda AI is building generalist intelligent robots with a full robotics stack from hardware to foundation world models. The company has raised over $450M and is scaling manufacturing and research aggressively.
You will own the end-to-end robot-learning pipeline, making it reliable, reproducible, and measurable from data collection through real-robot evaluation. This is a core research systems role requiring deep understanding of robot data semantics and model behavior—not generic DevOps or infrastructure support.
Key responsibilities include:
• Own the complete workflow: robot data collection, dataset generation, model training, inference, and real-robot evaluation.
• Define the golden path: maintain known-good combinations of code, datasets, configurations, checkpoints, robot software, hardware settings, and evaluation procedures.
• Build automated validation at every interface: timestamp synchronization, sensor integrity, episode completeness, schema compatibility, dataset migrations, dataloader outputs, preprocessing, model inputs, and configuration correctness.
• Create end-to-end regression tests: smoke tests for data compilation, training, checkpoint loading, inference, and real-robot execution. Develop small-scale overfit and canary experiments to catch regressions before expensive training runs.
• Ensure training-inference consistency: identify and prevent discrepancies in image processing, sensor normalization, temporal context, action representation, and model configuration.
• Make failures observable: build instrumentation and debugging tools to determine whether regressions originated in data, model code, infrastructure, inference, robot software, hardware, environment, or evaluation.
• Improve real-robot evaluation: partner with researchers and operations to establish stable benchmark stations, reference baselines, clear rubrics, repeatable protocols, and tracking of environmental variables.
• Lead cross-functional root-cause investigations: drive ambiguous failures to resolution across Research, Data Infrastructure, Model Infrastructure, Software, and Robot Operations.
• Establish release standards: define validation required before changes to robot software, data systems, training code, or inference systems.
• Measure research reliability: track pipeline success rates, reproducibility, regression frequency, time to root cause, and benchmark stability.
Required qualifications: strong track record owning or debugging complex ML, robotics, or autonomy systems across multiple layers. Excellent software engineering with Python, maintainable design, automated testing, CI/CD, and production debugging. Hands-on PyTorch or comparable ML framework experience. Experience building ML workflows spanning data, training, evaluation, and deployment. Strong systems debugging and ability to structure ambiguous problems. Familiarity with data-pipeline correctness, schemas, versioning, temporal data, migrations, provenance, and validation. Experience with Linux, containers, Kubernetes, Slurm, or distributed GPU clusters. Attention to detail for silent failures. Cross-functional collaboration without formal authority. Clear written communication.
