Helix AI Engineer, Android

Responsibilities

• Build and own the Android application that serves as the primary mobile interface to Figure's humanoid robots via USB Host / Android Open Accessory protocols.
• Architect high-throughput, zero-drop data ingestion pipelines for high-FPS image sensors and high-frequency IMU data using zero-copy memory techniques.
• Implement custom hardware abstraction layers (HAL) and leverage the Android NDK (C/C++) for high-performance, low-latency processing.
• Optimize CPU/GPU workloads for real-time edge filtering under strict thermal and battery constraints.
• Integrate on-device AI inference libraries such as TFLite, MediaPipe, ONNX Runtime, or OpenCV for real-time computer vision and sensor fusion.
• Implement low-latency video streaming protocols like WebRTC.

Requirements

• Deep expertise in Android NDK (C/C++) including custom HAL development and USB Host/AOA protocol communication.
• Proven experience architecting real-time, low-latency data pipelines for high-bandwidth sensors using zero-copy memory and real-time concurrency.
• Mastery of Android system resource management, including CPU/GPU optimization, thermal/battery constraints, and foreground services.
• Strong proficiency in both C/C++ (NDK) and Kotlin/Java.
• Experience shipping production Android applications in hardware-connected, latency-critical environments.
• Track record of maintaining production applications at scale, including crash rate management and OTA update strategies.

Preferred Qualifications

• Experience integrating on-device CV/ML inference (TensorFlow Lite, MediaPipe, ONNX Runtime, or OpenCV) to raw sensor feeds.
• Familiarity with WebRTC or other low-latency streaming protocols.
• Background in DSP techniques applied directly to raw sensor data.
• Prior work in robotics companion apps, industrial Android devices, AR/computer vision mobile apps, or drone control applications.

About the Company

Figure is an AI Robotics company developing a general purpose humanoid designed for commercial tasks and the home. We are based in San Jose and require 5 days/week in-office collaboration.