Matthew Wnuk

Matthew WnukStaff Embodied AI Engineer, Qualcomm

I build the models that make robots work in the physical world.

I work on foundation models for humanoids and robot arms: policies that see, predict, and act, and that run on the robot's own hardware. Before robots it was a decade of machine learning for cars, shelf cameras, warehouses, and radios.

See the work Resume LinkedIn GitHub Hugging Face Email
A Unitree G1 humanoid robot standing on a conference floor.
Unitree G1 humanoid

Robots I work with

A Unitree G1 humanoid robot on a show stage.
Unitree humanoidsFull-size and compact humanoids for whole-body control and manipulation.
A white Franka robot arm reaching over a table at a trade show.
FrankaThe seven-axis research arm behind much of today's manipulation data and benchmarks.
A silver Flexiv Rizon robot arm polishing a part.
Flexiv AICOA dual-arm mobile platform built on Flexiv's force-controlled Rizon arms (a Rizon arm is pictured).
A NEURA Robotics robot arm on a lit display stand.
NEURA LARANEURA Robotics' collaborative arm (its sibling MAiRA is pictured).

Photos from Wikimedia Commons: Unitree G1 by Sayanesy (CC0); Unitree on stage by RuinDig/Yuki Uchida (CC BY 4.0); Franka by Ims (CC BY-SA 4.0); Flexiv Rizon by MichelleACipriano (CC BY-SA 4.0); MAiRA by Anna Schneider2021 (CC BY-SA 4.0).

Selected work

Qualcomm · Staff Embodied AI Engineer · 2026 to present

World-Action Models for Robot Manipulation

Robot foundation models that predict future video and robot actions together, from multi-robot data to on-device control.

  • World-action models
  • Video generation
  • Imitation learning
  • On-device

Symbotic · Senior Software Engineer, ML and Perception · 2023 to 2026

CaseID: SKU Identification for Warehouse Automation

Identifying the SKU of every case a warehouse robot handles, with the whole pipeline running onboard the robot.

  • Vision Transformers
  • Metric learning
  • ONNX
  • NVIDIA Jetson
  • ROS

Modern Intelligence · Senior Machine Learning Engineer · 2022

Multimodal Retrieval for Maritime Awareness

Finding the same vessel across cameras, satellites, AIS, and text descriptions.

  • CLIP-style models
  • Re-identification
  • AIS
  • DVC
  • Weights and Biases

Sony Electronics · Staff Machine Learning Engineer · 2020 to 2022

Visual Speech Recognition (Lip Reading)

An in-car lip-reading system that keeps speech recognition working in noisy cabins.

  • 3D ResNets
  • Sequence-to-sequence
  • CTC
  • Patented

Sony Electronics · Staff Machine Learning Engineer · 2020 to 2022

Retail Applications

Shelf-mounted cameras that identify products and flag gaps in stock, running on the edge.

  • YOLOv5
  • Embeddings
  • TFLite
  • Edge inference

SPAWAR and DARPA · Machine Learning Engineer · 2017 to 2019

RFMLS: Radio Frequency Machine Learning Systems

Bringing deep learning to radio signals: a benchmark dataset, hardware fingerprinting, and learned receiver control.

  • Variational autoencoders
  • Complex-valued networks
  • Reinforcement learning

Experience

  1. 2026 to present

    Staff Embodied AI Engineer

    Qualcomm

    Foundation models for humanoid and general-purpose robots: vision-language-action policies and world-action models, and their deployment on Qualcomm edge hardware.

  2. 2023 to 2026

    Senior Software Engineer, ML and Perception

    Symbotic

    Tech lead for CaseID: image embedding with Vision Transformers for a large-scale production system, deployed on NVIDIA Jetson in an embedded ROS environment.

  3. 2022

    Senior Machine Learning Engineer

    Modern Intelligence

    A custom multimodal re-identification system for maritime awareness.

  4. 2020 to 2022

    Staff Machine Learning Engineer

    Sony Electronics

    Tech lead for visual speech recognition; senior engineer on embedded product recognition for retail.

  5. 2016 to 2020

    Machine Learning Engineer

    SPAWAR

    Lead engineer on DARPA's Radio Frequency Machine Learning Systems program; tech lead on machine learning for self defense; machine learning trainer.

  6. 2016 to 2018

    MS, Electrical Engineering

    UC San Diego

    Research on lidar SLAM and sensor fusion for an embedded drone platform.

  7. 2011 to 2015

    BS, Electrical Engineering

    San Diego State University

    President of Eta Kappa Nu; first-place team at the 2015 AUVSI RoboSub competition; internships at NASA Langley and Northrop Grumman.

  8. 2004 to 2013

    Sonar Technician

    United States Navy

    Nine years of service around the world.

Full resume →

Patents

Certifications and honors

  • NVIDIA-Certified: Generative AI LLMsDecember 2024
  • NVIDIA-Certified: Generative AI MultimodalDecember 2024
  • MIT Professional EducationModeling and optimization for machine learning; designing efficient deep learning systems · July 2019
  • Eta Kappa Nu and Tau Beta PiEngineering honor societies