Machine Learning Engineer, Perception

Wing•Palo Alto, CA

1d•Hybrid

About The Position

Wing is looking for a Machine Learning Engineer to join our Perception team. This role is hybrid based in Palo Alto, CA. You will be at the forefront of advancing our autonomous drone delivery system, leveraging deep learning to ensure our aircraft navigate the world with precision and safety. At Wing, you aren’t just building models in a vacuum; you have access to tens of thousands of hours of flight data from a diverse sensor suite. Your improvements will deploy directly to a large fleet of autonomous aircraft, impacting thousands of real-world customers daily. We value senior engineers who are intrinsically driven and have a knack for solving complex problems with stunningly simple, pragmatic solutions.

Requirements

3+ years of experience in ML, deep learning, or computer vision, ideally within robotics or autonomous systems.
MS degree in Computer Science, Robotics, Aerospace, Mechanical Engineering or 2 years of experience working on problems involving above mentioned skills in a directly related field
Excellent software design and implementation skills
Proficiency in Python and/or C++.
Hands-on experience with JAX, PyTorch, or TensorFlow, building training/inference/validation pipelines.
Familiarity with modern CV architectures (CNNs, Transformers, etc.).
Hands on experience in one or more of the following domains: Multi-view Geometry & 3D Reconstruction. Monocular or Stereo Depth Estimation. Object detection, localization, Semantic Segmentation. Local and global image features Vision foundation models, VLMs, VLAs
Ability to design and implement well thought-out simple solutions to business critical problems

Responsibilities

Build and deploy highly reliable ML solutions both for the backend as well as optimized for resource-constrained, real-time environments.
Support and improve existing ML models and infrastructure, sharing in the team’s expertise.
Partner across hardware, software, and flight-testing teams to take features from concept to production-ready flight code.