Lead Robotics Researcher, 3D Vision & World-Action Models

Sieve, Inc.; San Francisco (CA); On-site; USD 130,000 - 210,000; Full time

Benefits offered by this job

401(k) and full health insurance

Breakfast, lunch, and dinner covered,＋

Ubers covered home

Job summary

Sieve is seeking a Robotics Researcher in San Francisco to lead training and evaluation of world-action models and 3D perception systems. You will connect data capture, curation, and labeling with measurable gains in robot learning, running projects from question to answer in 3D-rich environments.

Establish experiments, evaluate on unseen tasks, and collaborate with partner labs. Expect a hands-on role in designing datasets, pipelines, and evaluation frameworks for real-world robotics data.

Qualifications

• 4+ years of research or engineering experience in robot learning, embodied AI, or 3D computer vision.
• Fluent Python plus PyTorch or JAX: you can write, debug, and change training code yourself.
• Hands-on 3D vision work: multi-view geometry, depth, pose, tracking, or reconstruction.
• Have trained a policy or world model and tested it on data it never saw in training.
• Rigorous about experiments: baselines, ablations, and attributing gains properly.
• Comfortable with large, messy datasets and diagnosing data quality issues.
• Strong communication skills and the ability to explain findings clearly.
• In-person at our SF HQ

Responsibilities

• Data ablations for robot learning: compare data sources and vary quality.
• World-action models: train and post-train action-conditioned models and policies.
• 3D labeling from video: build pipelines for depth, trajectories, pose, tracks, and 3D reconstruction.
• Evaluation design: lock test tasks and report clear conclusions.
• Simulation as a tool: use simulators or learned models for evaluation and augmentation.
• From findings to datasets: create curation filters and capture protocols.
• Partner collaboration: work with engineers and researchers to select problems.
• Publish what you learn: share findings via papers, blogs, or open data

Skills

Python

PyTorch

JAX

3D vision

Robotics

Tools

MuJoCo

Isaac Sim

Job description

Sieve is seeking a Robotics Researcher in San Francisco to lead training and evaluation of world-action models and 3D perception systems. You will connect data capture, curation, and labeling with measurable gains in robot learning, running projects from question to answer in 3D-rich environments.

Establish experiments, evaluate on unseen tasks, and collaborate with partner labs. Expect a hands-on role in designing datasets, pipelines, and evaluation frameworks for real-world robotics data.