Jobs at Rhoda AI

Design, build, and scale high-throughput data infrastructure for multimodal training data. Optimize storage, indexing, retrieval, throughput, and workload balancing. Implement observability, artifact versioning, and tooling to enable research and production use. Integrate VLMs and support reproducible, cost-efficient pipelines for large-scale model training.

Develop and implement post-training pipelines that adapt web-pretrained causal video models to real robot tasks. Build RL training pipelines, reward models, inverse dynamics for action execution, and evaluation frameworks. Iterate between simulation and hardware, diagnose failures, and improve generalization with few-shot adaptation and preference alignment on robot-collected demonstration data.

Optimize inference performance of large multimodal foundation models across cloud and on-robot targets. Diagnose bottlenecks, apply quantization/pruning/distillation, tune kernels (CUDA/Triton), build benchmarking and regression detection, and translate research models into deployment-ready implementations.

Research and implement model-compression and efficiency techniques (quantization, pruning, distillation, low-rank approximations), design efficient architectures and attention mechanisms, develop training strategies, profile and benchmark models on hardware, build evaluation frameworks, and collaborate with training and deployment teams to enable real-time inference on robot hardware. Publish and present results.

Build internal frontend interfaces, backend services, and tooling connecting cloud, data pipelines, and robot fleets. Own observability, scalability, and reliability of platform tools, collaborate with researchers and engineers, and integrate storage, APIs, and deployment/teleoperation systems.

Develop and maintain real-time perception pipelines for humanoid robots: sensor integration and calibration, object detection/tracking, low-latency streaming to models and teleop, safety-certified perception, field bring-up, and reliability/fault recovery.

Design, prototype, and integrate mechanical components for a Gen 0 humanoid robot. Work includes CAD modeling, BOM creation, end effector and mount design, third-party component integration, QA testing, and root-cause analysis. Collaborate with cross-functional engineers and support manufacturing handoffs with a focus on safety, reliability, and manufacturability.

Research and develop learning-based methods for contact-rich, dexterous manipulation: design training/data collection, perception (contact/tactile/pose), build generalizable policies, create simulation benchmarks, and close sim-to-real with hardware teams.

Lead design and implementation of teleoperation and demonstration collection systems, build curation and quality pipelines, source and augment robotic datasets, create tooling for annotation and exploration, and measure downstream impacts on model and policy performance.

Lead design and implementation of large-scale multimodal training systems. Diagnose and optimize compute, communication, and memory bottlenecks across thousands of GPUs; define parallelism strategies; build observability and regression detection tools; collaborate with researchers and infra to improve distributed efficiency and scaling for robotics world models.

Own and develop low-level real-time firmware for distributed humanoid robot control over EtherCAT. Design deterministic communication and synchronization (<1ms), implement safety mechanisms, tune RTOS and DMA/ISR paths, configure EtherCAT masters/slaves, collaborate across hardware and controls teams, and develop unit/HIL/integration tests and documentation for firmware validation and bring-up.

Lead architecture and delivery of multi-sensor imaging and perception systems: select and integrate cameras, ToF, radar, LiDAR, and thermal sensors; own camera signal chain and ISP tuning; design illumination and sensor fusion hardware; ensure signal integrity, EMI/EMC, and production readiness through EVT/DVT/PVT to mass production; collaborate cross-functionally and mentor engineers.

Lead full-cycle HR for a fast-growing robotics startup: manage onboarding/offboarding, HRIS integrity, benefits and compensation, performance and L&D programs, culture and engagement, compliance and multi-state employment law, and coordinate visa sponsorship and immigration processes.

Design, build, and maintain automated test plans, fixtures, and infrastructure across unit, integration, system, and hardware-in-the-loop layers. Extend CI/CD pipelines, define quality metrics and KPIs, track coverage and failures, and participate in design/code reviews to improve observability, testability, and reliability of the robotics software stack.

Lead hardware engineering to design, develop, validate, and scale robotic systems from concept to volume production. Define architecture, oversee mechanical/electrical/system design, test and certification, NPI, supplier relationships, and build a high-performing hardware organization focused on reliability, safety, and manufacturability.

Work with customers to translate requirements into model adaptation strategies, fine-tune foundation world models for domain tasks, design experiments and benchmarks, validate behavior in real environments, and feed deployment insights back to core research teams.

Research and develop multi-step reasoning and planning methods for embodied world models; design architectures and training strategies for compositional generalization and long-horizon prediction; explore chain-of-thought, process reward models, and test-time search; build evaluation benchmarks; collaborate across pre/post-training and robotics teams; publish research.

Build and operate cloud infrastructure powering data collection, robot field operations, and model training. Ensure reliability, low latency, and scalability across databases, object storage, compute, and backend services; develop observability, debug production incidents, and collaborate with research and robotics teams.

Build and maintain large-scale training orchestration and experiment management systems for distributed GPU model training. Implement observability, scheduling, artifact management, and tooling to optimize research iteration, cluster utilization, and reliability while collaborating with research and infra teams.

Build and maintain onboard networking, data pipelines, telemetry, tooling, and perception-control integration for humanoid robots. Support hardware bring-up, field testing, and cross-team collaboration to ensure reliable data collection and deployment.