Robotics engineering · Detroit

Robotics, perception, and applied AI engineering.

Dexteris AI is your home-grown robotics and enterprise systems engineering consultants. Founded and based in Michigan, Dexteris AI was built to create real, useful results from AI integrations in robotics, automation, and enterprise systems. We build solutions for manipulation, mobility, vision, and industrial systems - along with the AI learning pipelines that ensure solutions stay relevant to your changing needs.

Discuss a project Selected work
Two-link inverse kinematics · click to aim auto

Selected work

Projects

2026 · Venture studio, Detroit

Humanoid teleoperation and manipulation platform

An operator wearing a Meta Quest headset drives the robot's upper body directly, in real time. A web application gives people without a headset the same control from a browser. Obstacle avoidance and a whole-body control adaptor run underneath both, so neither operator has to think about the robot's limits.

Behind the teleoperation sits a reinforcement-learning training pipeline that produces manipulation policies and carries them across to the physical robot. Transfer is verified on hardware.

Platform
Unitree G1
Delivered
XR teleoperation, browser-based control application, obstacle avoidance, whole-body control adaptor, reinforcement-learning training pipeline with verified sim-to-real transfer
A Unitree G1 humanoid robot standing on an exhibition hall floor, its visor lit blue.
Unitree G1

Capabilities

What we work on

These overlap in practice. A gripper that misses is usually a perception problem, and a vision system that drifts is usually a mechanical one, so most engagements draw on two or three of them at once.

Robotics & motion

  • Hardware diagnostics and bring-up
  • SDK and driver integration
  • Whole-body control
  • Motion planning
  • Teleoperation and XR interfaces
  • ROS 2 integration

Vision & perception

  • 3D perception and reconstruction
  • 6-DoF pose estimation
  • Detection and segmentation
  • Hand-eye calibration
  • Sensor fusion
  • Obstacle avoidance

Applied AI

  • Reinforcement learning
  • Sim-to-real transfer
  • Learning from demonstration
  • Data collection pipelines
  • Policy evaluation
  • Task and skill orchestration

Industrial systems

  • Robot cell integration
  • PLC and line interfacing
  • Machine vision inspection
  • Safety interlocks and watchdogs
  • Operator interfaces
  • Commissioning and training

Stack

Platforms and tools

Robots

  • Humanoid platforms
  • 6/7-DoF arms
  • Mobile bases

Sensing

  • Meta Quest
  • Depth cameras
  • IMU

Learning

  • PyTorch
  • Reinforcement learning
  • Sim-to-real

Software

  • ROS 2
  • Python
  • C++
  • Docker
  • TypeScript
  • Linux

Team

Who you'll work with

You work directly with the engineer writing the code. Strong Academia ties guarantee cutting edge solutions.

Blaine Oania

Principal engineer

Blaine works where software touches the real world. Specializing in Embodied AI and enterprise industrial systems, his background spans controls design, industrial software, and robotics.

  • M.S. Robotics and Intelligent Control, Wayne State University, 2026 · B.S. Computer Science, 2023
  • Michigan DevFest 2024 “How to Train Your Robot!”
  • Robotics, Automation, Enterprise Systems, Controls

Contact

Start with a conversation

Tell us what the equipment is and what you need it to do. We'll tell you whether it's something we can build, what it realistically takes, and what we'd want to prove first. No charge for that, and no obligation either way.

If it's a fit, we'll put a scope in front of you.

blaine@dexterisai.com LinkedIn Detroit, MI · on-site with hardware, remote for everything else
Opens in your mail client. Nothing is sent from this page.