Senior Machine Learning Engineer - Robotics

Johnson & Johnson Johnson & Johnson · Pharma · Santa Clara, CA +1

Senior Machine Learning Engineer focused on applying learning-based manipulation to improve surgical robotics performance and safety. The role involves designing models, implementing algorithms, and deploying ML-trained behaviors from research to real-world hardware, with responsibilities including mentoring and elevating team technical capabilities. Key tasks involve developing ML algorithms for task automation, building data pipelines, leveraging simulation for training and sim-to-real transfer, defining performance metrics, collaborating on deployment, and optimizing models for real-time performance.

What you'd actually do

  1. Developing and implementing ML algorithms for surgical robotic task automation.
  2. Building pipelines for data collection, training, testing and deployment of ML models.
  3. Leverage simulation for prototyping, training, and transfer to robot hardware.
  4. Define and evaluate performance metrics for surgical tasks and implement rigorous testing protocols.
  5. Collaborate with software and controls teams to deploy learned policies and ensure safe and reliable execution.

Skills

Required

  • Machine Learning Algorithms
  • Python (Programming Language)
  • Robotics
  • Python for prototyping
  • C++ for deployment on robotic hardware
  • deep learning frameworks (e.g., PyTorch, TensorFlow)
  • optimizing ML models for real-time performance on robotic hardware
  • simulation platforms (e.g., MuJoCo, Isaac Sim)
  • sim-to-real transfer
  • robot kinematics, dynamics, sensing, and control

Nice to have

  • Critical Thinking
  • Dev-C++
  • Industry Analysis
  • Innovation
  • Problem Solving
  • Process Improvements
  • Prototyping
  • Research and Development
  • Robotic Automation
  • Robotic Control Software
  • Vision-Language-Action Models (VLAs) for Robots
  • build tools and version control systems
  • agile development processes
  • CI/CD
  • issue tracking tools

What the JD emphasized

  • ML-trained manipulation behaviors
  • real-world robotic hardware
  • large multimodal ML models
  • onboard real-time performance on robotic hardware
  • sim-to-real transfer

Other signals

  • ML-trained manipulation behaviors
  • learning-based manipulation
  • ML models for surgical robotic task automation
  • learned policies