ML Infrastructure Engineer

Sunday•Mountain View, CA

About The Position

At Sunday, we're developing personal robots to reclaim the hours lost to repetitive tasks. We're focused on an ambitious goal to make generalized robots broadly accessible, enabling households to take back quality time. We have spent the last 18 months building a talented team, securing capital, and validating our technology. We are now seeking passionate individuals to join us in the next phase of our growth. If you are ready to apply your skills to the forefront of robotics innovation, we’d love to hear from you. Sunday Robotics is building the future of home robotics. We're developing end-to-end ML models for robot manipulation, and you'll have the opportunity to build and shape foundational systems that directly accelerate our path to putting robots in homes. This is a broad role that can be tailored to your specific area of expertise: data pipelines, training infrastructure or inference. You'll build systems across the full robot learning pipeline: ingesting and processing multimodal data, scaling distributed training, optimizing inference for real-time control and building research tooling.

Requirements

Strong software engineering and systems fundamentals
Experience building distributed systems or large-scale data pipelines
Hands-on experience with ML training infrastructure, ideally PyTorch
Comfort reasoning about performance, memory, I/O, and GPU utilization
Experience managing training workloads (SLURM, Kubernetes, or similar)
Ownership mindset: you design, build, operate, and iterate on systems end-to-end
Enjoy working closely with researchers and unblocking fast-moving projects

Nice To Haves

Experience with robotics data pipelines or multimodal models
Background in VLAs, Video Generation architectures or robot learning systems
Deep ML systems experience: training compilers, custom kernels, runtime optimization
Hands-on GPU performance tuning
Experience with serialization formats for high-performance systems (Protobuf, FlatBuffers, MCAP)

Responsibilities

Maintain an effective research codebase with good ergonomics, optimizing for fast iteration and correctness
Own infrastructure for model training: job scheduling, checkpointing, metrics, and logging
Scale distributed training across GPU clusters with minimal researcher friction
Enable training of larger models through sharding, activation checkpointing and memory optimization
Profile and optimize gpu utilization, memory usage and training throughput
Build low-latency inference pipeline for real-time robot control, apply quantization, distillation and model compilation to optimize inference performance
Work closely with researchers and roboticists to translate research needs into reliable software and infrastructure
Design high-throughput pipelines for ingesting, validating, and transforming multimodal robot data (video, proprioception, actions)
Build storage systems and metadata indexing for efficient dataset management at large scale
Optimize dataloaders, sharding and prefetching to minimize time from data arrival to model training
Build research tooling for debugging, visualization and experiment analysis