Robotics Engineer

Auxo Talent; New York (NY); On-site; USD 140,000 - 210,000; Full time

Benefits offered by this job

Founding equity

Job summary

Auxo Talent is seeking a founding engineer to lead perception and robot learning for a warehouse robotics startup. You will set the technical direction, design approaches, and build with the founding team on live deployment.

You will work with a small, fast-moving team, shipping models to customer operations within months, and expanding scope with every new robot and task. Equity is provided.

Qualifications

• 3–5 years hands-on robotics, CV or ML experience, ideally with physical robots.
• Experience training models on GPUs and deploying where speed matters.
• 3D perception with point clouds and depth cameras.
• Experience with real robot hardware and simulation (MuJoCo, Isaac Sim or similar).
• Strong Python, plus C++ or another systems language.
• Prototyping quickly and iterating on demos.

Responsibilities

• Lead perception and robot learning to set technical direction.
• Develop and deploy models on live robots within constraints.
• Advance data pipelines from raw recordings to training data.
• Scale models across new robots and customers.

Skills

Python

C++

3D perception

GPU training

ML

Robotics

Tools

MuJoCo

Isaac Sim

Job description

This role will require substantial travel to different customers' sites and warehouses.

About the company

We're supporting a seed-stage warehouse robotics startup. Their software is already running in large enterprise operations. Now they're building the AI that lets each robot recognize products and handle them on its own. The founders have built and scaled robotics teams before.

The role

This is a founding engineering role. You'll lead perception and robot learning and make the key technical decisions. There's no large legacy codebase, so you'll design the approach yourself and build it with the founding team. Success means models that work reliably on live warehouse floors.

What you'll own

• Perception: 3D scene understanding with depth cameras and point clouds, object pose estimation, and identifying new products without retraining
• Robot learning: Training and fine-tuning vision-language-action models, using simulation and synthetic data
• Deployment: Running models on real robots within tight speed and compute limits, and feeding real-world mistakes back into training
• Data: Teleoperation, data collection, and the pipeline that turns raw recordings into clean training data

Why it's worth a look

• You'll set the technical direction for perception and learning
• Your models get tested in real customer operations within months
• A small team that makes decisions quickly
• Your scope grows with every new robot, task and customer
• Founding equity

What you'll bring

• 3 to 5 years of hands-on work in robotics, computer vision or ML, ideally on physical robots
• Experience training models on GPUs and deploying them where speed matters
• 3D perception experience with point clouds and depth cameras
• Experience with real robot hardware and simulation (MuJoCo, Isaac Sim or similar)
• Strong Python, plus C++ or another systems language
• You like getting prototypes working quickly

Nice to have

• Training manipulation or vision-language-action policies
• Grasp planning
• Warehouse or logistics automation experience