Senior Perception Learning Engineer - SLAM

Apptronik; Sunnyvale (CA); On-site; USD 190,000 - 235,000; Full time

Job summary

Apptronik is a human-centered robotics company building AI-powered robots to assist in manufacturing, logistics, and beyond. This role drives R&D of perception systems for Apollo, focusing on SLAM, VIO, world modeling, and learning-based perception to enable robust autonomy in real-world settings.

You will design and optimize deep learning models for real-time detection, tracking, segmentation, scene understanding, and state estimation while integrating multi-modal data into unified world models

Qualifications

• MS/PhD in Computer Science, Robotics, Computer Engineering, or related field.
• 3–5+ years building and shipping perception and/or SLAM systems in robotics or real-time vision applications.
• Strong background in SLAM, VIO, probabilistic state estimation, and deep learning for computer vision.
• Hands-on experience with modern AI frameworks (PyTorch, JAX, TensorFlow) and CV / multi-modal libraries such as OpenCV, Detectron2, YOLO, and foundation models for perception and language (e.g., SAM, CLIP, DINOv2, Flamingo).
• Proficiency in Python and modern C++, with strong software engineering fundamentals (version control, testing, CI/CD).
• Deep understanding of 3D geometry, camera models, and probabilistic estimation (EKF/UKF, factor graphs, SLAM, VIO).
• Experience deploying optimized perception or SLAM models on edge hardware (GPU/NPU/embedded) under compute, latency, and thermal constraints.
• Track record of shipping ML/Perception or SLAM systems from R&D into production robotics platforms.

Responsibilities

• Lead the design, development, and optimization of perception and SLAM pipelines for humanoid robots, including visual-inertial odometry, mapping, localization, object detection, tracking, segmentation, pose estimation, and scene understanding.
• Develop multi-sensor fusion frameworks integrating cameras, LiDAR, depth sensors, and IMUs for robust real-time state estimation and mapping in dynamic, human-centered environments.
• Contribute to scalable data pipelines, training infrastructure, and inference frameworks to accelerate model development, evaluation, and deployment.
• Drive research and deployment of learning-based models for SLAM, 3D scene understanding, and perception optimized for humanoid locomotion, manipulation, and human-robot interaction.
• Implement performance profiling, regression testing, and telemetry systems to ensure perception and SLAM modules meet strict latency, accuracy, and reliability requirements on edge devices.
• Collaborate with planning, control, and hardware teams to define perception-to-action interfaces, ensuring real-time compatibility with locomotion and manipulation pipelines.
• Guide the integration of synthetic data (e.g., IsaacSim) with real-world datasets to enhance model generalization and robustness.
• Contribute to best practices in code quality, model versioning, reproducibility, and deployment.

Skills

SLAM

VIO

World modeling

Learning-based perception

Object detection

Multi-sensor fusion

Education

MS/PhD in CS/Robotics

Tools

PyTorch

JAX

TensorFlow

OpenCV

Detectron2

YOLO

Job description

Apptronik is a human-centered robotics company developing AI-powered robots to support humanity in every facet of life. Our flagship humanoid robot, Apollo, is built to collaborate thoughtfully with people, starting with critical industries such as manufacturing and logistics, with future applications in healthcare, the home, and beyond.

We operate at the cutting edge of Applied AI, applying our expertise across the full robotics stack to solve some of society's most important problems. You will join a team dedicated to bringing Apollo to market at scale, tackling the complex challenges like safety, commercialization, and mass production to change the world for the better.

Job Summary

You will drive research and development of advanced perception systems that empower Apptronik's humanoid robots to understand and interact with complex human environments. Your work will focus heavily on SLAM, visual-inertial odometry, world modeling, and learning-based perception, alongside object detection and multi-sensor fusion, creating the foundation for robust autonomy in real-world settings.

You will design and optimize deep learning models for real-time detection, tracking, segmentation, scene understanding, and state estimation while contributing to scalable pipelines for training, evaluation, and deployment. You will also integrate data from multiple modalities — cameras, LiDAR, depth sensors, and IMUs — into unified world models that support navigation, manipulation, safety, and human-robot interaction.

This role requires balancing research innovation with practical engineering to deliver deployable, high-performance SLAM and perception stacks. You will collaborate across Reinforcement Learning, Platform Software, and Systems teams, and contribute to shaping Apptronik's perception and autonomy roadmap. Your work will directly accelerate the development of humanoid robots that can safely operate in human spaces, adapt to dynamic environments, and extend human capability.

ESSENTIAL DUTIES AND RESPONSIBILITIES Or KEY ACCOUNTABILITIES

• Lead the design, development, and optimization of perception and SLAM pipelines for humanoid robots, including visual-inertial odometry, mapping, localization, object detection, tracking, segmentation, pose estimation, and scene understanding.
• Develop multi-sensor fusion frameworks integrating cameras, LiDAR, depth sensors, and IMUs for robust real-time state estimation and mapping in dynamic, human-centered environments.
• Contribute to scalable data pipelines, training infrastructure, and inference frameworks to accelerate model development, evaluation, and deployment.
• Drive research and deployment of learning-based models for SLAM, 3D scene understanding, and perception optimized for humanoid locomotion, manipulation, and human-robot interaction.
• Implement performance profiling, regression testing, and telemetry systems to ensure perception and SLAM modules meet strict latency, accuracy, and reliability requirements on edge devices.
• Collaborate with planning, control, and hardware teams to define perception-to-action interfaces, ensuring real-time compatibility with locomotion and manipulation pipelines.
• Guide the integration of synthetic data (e.g., IsaacSim) with real-world datasets to enhance model generalization and robustness.
• Contribute to best practices in code quality, model versioning, reproducibility, and deployment.

EDUCATION And/or EXPERIENCE

• MS/PhD in Computer Science, Robotics, Computer Engineering, or related field.
• 3–5+ years building and shipping perception and/or SLAM systems in robotics or real-time vision applications.
• Strong background in SLAM, VIO, probabilistic state estimation, and deep learning for computer vision.
• Practical expertise in detection, segmentation, multi-object tracking, 3D perception, and learning-based mapping/localization.
• Hands-on experience with modern AI frameworks (PyTorch, JAX, TensorFlow) and CV / multi-modal libraries such as OpenCV, Detectron2, YOLO, and foundation models for perception and language (e.g., SAM, CLIP, DINOv2, Flamingo).
• Proficiency in Python and modern C++, with strong software engineering fundamentals (version control, testing, CI/CD).
• Deep understanding of 3D geometry, camera models, and probabilistic estimation (EKF/UKF, factor graphs, SLAM, VIO).
• Experience deploying optimized perception or SLAM models on edge hardware (GPU/NPU/embedded) under compute, latency, and thermal constraints.
• Track record of shipping ML/Perception or SLAM systems from R&D into production robotics platforms.

Preferred Qualifications

• Experience with humanoid robots, bipedal locomotion, and manipulation tasks.
• Experience with modern SLAM frameworks (e.g., ORB-SLAM, VINS, Rtabmap, GTSLAM, Cartographer) or learning-based SLAM systems.
• Strong classical computer vision skills (geometry-based methods, feature extraction) complementing learning-based approaches.
• Expertise in model acceleration, quantization, or compression (TensorRT, ONNX Runtime).
• Familiarity with real-time frameworks and middleware such as ROS 2, GStreamer, or zero-copy pipelines.
• Knowledge of synthetic data generation and domain adaptation techniques for training perception and SLAM models.
• Contributions to open-source robotics or vision software stacks.

Physical Requirements

• Prolonged periods of sitting at a desk and working on a computer
• Must be able to lift 15 pounds at times
• Vision to read printed materials and a computer screen
• Hearing and speech to communicate

The annual salary range is $190,000 - $235,000

Apptronik provides equal employment opportunities to all employees and applicants for employment and prohibits discrimination and harassment of any type without regard to race, color, religion, age, sex, national origin, disability status, genetics, protected veteran status, sexual orientation, gender identity or expression, or any other characteristic protected by federal, state or local laws.