Tom Donahue

Currently in stealth mode, Familiar Machines & Magic is creating artificial life to build a more caring world. Our EQ-first Physical AI platform blends personality, memory, and motion so machines can relate, remember, and support their users.

Focused on imbuing character and social intelligence into Astro, Amazon’s mobile, autonomous, home robot. • Designed and built core architectural components for intelligent proactivity and planning via LLM agent hybrid architecture • Led design, development and deployment of attention system which provides Astro with intentional gaze behaviors and real-time reactions to environmental stimuli Languages: Java, Python Technologies: Android, AWS

Engineering Lead in charge of architecting, developing and evaluating cutting-edge conversational technologies and character experiences for Jibo. Languages: Typescript / JS / Node Technologies: Azure Speech / Cognitive Services, Neo4j, Microsoft BotFramework

Member of the team that investigates, designs, implements and integrates high-level behaviors into gita and beyond, focusing primarily on enabling and ensuring reliable, intuitive, and seamless human-robot interactions in every situation. Languages: C++ (17), Python Technologies: ZeroMQ, ROS, Realsense, Robotics, Ubuntu, Git

Member of the Character AI team, tasked building the subsystems and tools necessary to make the first social robot for the home—Jibo—feel lifelike, produce dynamic behavior, and ensure a consistent character experience across his wide range of interactions and skills. Languages: Typescript/JS, Python Technologies: Electron

Member of a small, agile team, as well as apart of a larger multinational team with others headquartered in Paris and Tokyo. I designed and built immersive behavior and functionality into Aldebaran's range of autonomous humanoid robots (Pepper and NAO). I worked on everything from sensor extraction to consumer apps, NLP to front-end web and much in-between. Languages: Python, JS, HTML/CSS, C++

Human-Robot Interaction and AI research; heavily focused on natural language understanding and generation. Languages: Java, JS, PHP, HTML/CSS, R, LaTeX

Human-Computer Interaction research, with strong emphasis placed on tangible interfaces and accessible computing. Languages: Java, C++, HTML/CSS

(Now known as the Timing, Attention & Perception Laboratory at Michigan State University.) Built a neural-network model of auditory perceptual categorization. Resulting poster was accepted and presented at the National Conference on Undergraduate Research. Languages: MATLAB

Withdrew in good standing from a Computer Science and Cognitive Science joint-Ph.D. program.

Thesis: What's closeness got to do with it? Investigating the effects of interface closeness on abstract problem solving and learning.

Mathematics Minor