Inventor & Systems Architect

Kevin Lee
Masters

Building the infrastructure that democratizes creative expression

I architect distributed systems, engineer deterministic inference engines, and pioneer psycho-acoustic synthesis methods for the next generation of embodied AI.

Explore Work

The Architect

Systems Engineering

Architecture & Design

Building distributed systems with deterministic data forensics, fail-closed guardrails, and bounded agent autonomy. The PRISM AI architecture spans embodied reasoning, psycho-acoustic synthesis, and cinematic virtual production.

Patent Inventor

US Patent Application

PRISM AI introduces a tri-pillared system combining volumetric laser-galvo rendering, real-time LLM behavior generation, and on-set AI embodiment for film, television, and live events.

Audio Engineering

MÜZI Platform

Pioneered simulated vocal physicality for AI voice synthesis—resolving continuation anomalies through respiratory cadences. Emotional cognition architecture with targeted musical arrangement generation.

Forensic Engineering

BlueBird System

Deterministic inference engine for employment law litigation—same input, same verified output, every time. Daubert/FRE compliance gates with Merkle tree verification and OCR processing.

PRISM AI

US Patent Application | Filed January 2026

Film and Media Interactive & Reactive AI Collaborator

A system and method for manifesting a generative-AI agent as a dynamic, embodied collaborator across film, television, live events, theme parks, and medical training.

"No prior art—patent or non-patent—teaches the tri-pillared combination of volumetric laser-haze rendering, on-the-fly LLM behavior generation, and on-set AI embodiment."

Volumetric Laser-Galvo

High-speed galvanometer scanners sweep laser beams through controlled haze volume, creating floating 3D light strokes and volumetric shapes.

Real-Time AI Core

Transformer-based LLM plus vision-perception pipeline processing skeletal, gesture, and voice streams to output behavior intents in under 100ms.

On-Set Engine

ROS2-based middleware orchestrating millisecond-scale pub/sub among sensors, AI, projection, lasers, and ambisonic audio.

Systems Built

BlueBird

Deterministic Forensic Data Engine

A deterministic inference engine for employment law litigation—same input, same verified output, every time. Built with Tauri/Rust, Python guardrail core, Merkle trees, and Daubert/FRE compliance gates.

Rust Python Daubert/FRE Merkle Trees OCR

MÜZI

Psycho-Acoustic Synthesis Core

An emotionally aware audio synthesis environment. Pioneered simulated vocal physicality—resolving AI voice continuation anomalies by injecting natural respiratory cadences.

Audio Synthesis Voice Cloning Respiratory Modeling Sentiment Analysis

NODE

Embodied Reasoning & Autonomy Layer

System for grounding AI decisions in physical reality through sensor fusion and real-world interaction frameworks. Enables embodied AI to perceive, reason, and act within spatial contexts.

Sensor Fusion Spatial Reasoning Embodied AI

PRISM AI

Master Ecosystem

Distributed local-first architecture unifying deterministic data forensics, emotionally aware audio synthesis, and cinematic virtual production pipelines. Fail-closed guardrail systems.

Local-First Guardrails Multi-Modal Data Sovereignty

MÜZI: CONSTELLATIONS

Production-ready tracks generated through MÜZI's psycho-acoustic synthesis pipeline. Proof of concept for emotionally-aware AI music generation.

Listen on SoundCloud

Left Me In The Rain

Released May 2026

The Star

Released March 2026

To Need Me

Released March 2026

Grey Matter Stains

Released February 2026

Education

Vocal Performance

Jacobs School of Music

Indiana University
2015-2020

Art & Design

Eskanazi School

Indiana University
2015-2020

Music Technology

IUPUI

President, Mu Tau Psi
2013-2015

Musical Theatre

AMDA

NYC
2002-2003

Let's Connect

Seeking senior technical positions in AI product development, embodied AI systems, audio engineering, or creative technology leadership. Remote strongly preferred.