Perception
Systems that interpret real environments: object and defect detection, pose estimation, depth and LiDAR sensing, and tracking that stays registered to physical hardware.
AI-Powered Engine Quality ControlAI AGENTS · COMPUTER VISION · SPATIAL COMPUTING
I build AI systems that understand and act in the physical world.
I combine AI agents, computer vision, spatial computing, and sensor data to create working tools for complex real-world workflows.
CH.01
Mixed Reality
Headset experiences that guide people through physical tasks.
Medical training
HoloLens 2 Knee Replacement Surgery Guide
Mixed reality training app that walks medical students through total knee replacement in 3D, adapting to their decisions in real time.
10 min saved per training session
Case studyAerospace maintenance training
Mixed Reality Naval Maintenance Trainer
Gaze-driven HoloLens 2 trainer for F-18 maintenance procedures, with an agent pipeline that flags struggling trainees to supervisors.
+15% task accuracy / -20% training time
Case studyAerospace maintenance
AI-Powered Aircraft Inspection
XR inspection system that spots panel defects in real time and auto-generates the maintenance documentation that normally follows by hand.
-25% documentation overhead
Case studyAutomotive XR
3D Engine Bay Scanning & Tracking
Full pipeline for turning a physical engine bay into a tracked XR experience: scan, optimize, register, and overlay on the real hardware.
Case studyIndustrial IoT
Live IoT Data in Mixed Reality
HoloLens 2 app that pairs a virtual drone with live MQTT sensor streams, showing environmental telemetry beside the physical asset.
Case studyCH.02
Vision on the line
Real-time detection that gives operators an instant verdict.
Manufacturing
AI-Powered Engine Quality Control
Real-time inspection system that catches missing and incorrectly installed engine components on the assembly line before they ship.
Case studyManufacturing
AI Spark Plug Inspector
Real-time defect detector that marks each spark plug pass or fail on screen, giving line operators an instant verdict.
Case studySmart city
AI Traffic Analysis
Vehicle counting and flow analysis that tracks traffic through configurable detection zones to surface density patterns in real time.
Case studyCH.03
Field & data
Perception taken outdoors, and the data that trains it.
Agriculture
AI-Powered Livestock Measurement
Non-contact cattle measurement on an iPad, combining pose estimation with LiDAR depth to compute real-world body dimensions.
Case studyML tooling
Unity Synthetic Data
Synthetic training data generated in Unity Perception to build labelled computer-vision datasets without hand-annotating real footage.
Case studyAbout
Allen Cass
I build multimodal AI, computer-vision, XR, mobile, and edge systems that connect digital intelligence to physical work. Most of my projects start with someone doing a difficult job by hand - inspecting an aircraft panel, rehearsing a surgical procedure, measuring livestock - and end with a working system that sees the task, reasons about it, and produces something useful. I design and build the full stack: model, interaction, application, device integration, and deployment.
- Ships into US Navy and USDA programs
- 10+ years across XR, computer vision, and agent systems
- Prototype to deployed system, solo or lead
What I build, end to end
Reasoning & Agents
Multi-agent pipelines that turn sensor and process data into decisions: tool use, retrieval over technical documentation, and automated reporting into live systems.
AI-Powered Aircraft InspectionSpatial & Interactive Systems
Headset and mobile experiences that guide people through physical tasks: gaze and hand interaction, 3D scanning, spatial registration, and in-world data visualisation.
HoloLens 2 Knee Replacement Surgery GuideProduction Engineering
The rest of the system: application code, APIs, databases, edge deployment on constrained hardware, and the integration work that makes a prototype usable.
Mixed Reality Naval Maintenance TrainerExperience
Senior AI Agent Engineer · ThreadX
- System Architecture: Architected a multimodal, AI-powered LMS bridging web, Meta Quest (MR), and RealWear (AR) ecosystems.
- Web Platform: Built the core LMS with Flutter and Firebase (Firestore, Auth, Functions), integrating a Gemini AI chat interface.
- AI Agentic RAG: Engineered a backend retrieval pipeline using Qdrant vector databases for intelligent, context-aware training recommendations.
- XR Simulations: Developed Unity-based immersive training using the XR Interaction Toolkit, syncing user progress seamlessly from web to headset.
- Industrial Wearables: Created a hands-free, voice-optimized companion app for the RealWear Navigator 520 for procedural maintenance.
- In-Headset AI: Embedded a natural-language AI assistant directly within the XR applications for real-time trainee support.
Innovation Lead - AI Agent Developer AI/AR SW Engineer III · Avatar Partners
- V-22 XR Maintenance Application - System: Engineered an N8N and PostgreSQL agentic workflow integrating with a HoloLens 2 application for V-22 aircraft. Impact: Automated defect triage, queried manuals via RAG, and updated logs in real-time, reducing documentation overhead by 25%.
- F-18 AI-Augmented Training - System: Led development of an XR maintenance solution using N8N and LLM agents to process xAPI data from a Learning Record Store (LRS). Impact: Automated error pattern analysis for corrosion detection and alerted supervisors, increasing task accuracy by 15% and cutting training time by 20%.
- ARSurgeon Cognitive Assistant - System: Spearheaded the integration of a Claude API agent into a HoloLens 2 / MRTK knee replacement training application. Impact: Delivered real-time procedural guidance via custom knowledge bases, reducing training time by 10 minutes per session.
- USDA Livestock Monitoring - System: Developed an iPad LiDAR and Unity3D app using YOLOv11 pose estimation and OpenCV for real-time distance tracking. Impact: Streamlined data pipelines using CVAT, Unity Perception (synthetic data), and AI agents for automated batch processing and reporting.
AR/XR/AI Mobile Software Developer · Self Employed - Freelance
- Led cross-platform mobile development using Unity3D for AR/VR/XR applications.
- Architected data-driven AR platform using AWS cloud infrastructure and AI services.
- Built interactive AR chatbot experience using Google AI Dialogflow.
- Managed distributed team of developers and established best practices.
Sr. Software Engineer AR/XR/AI/ML · MERIDIUN
- Led development of medical training application for Edwards Lifesciences.
- Architected Walmart ExplorAR, an iOS based AR HR benefits application.
- Developed Ford Escape AR Experience prototype for virtual vehicle interaction.
- Co-invented patent-pending AR/NFC anti-counterfeiting technology.
Certifications
- 2025 Generative AI Leader · Google
- 2025 AI Agents Fundamentals · Hugging Face
- 2025 Generative AI Certification · Kaggle / Google
- 2023 Microsoft Certified: Azure AI Fundamentals · Microsoft
- 2022 Deepstream For Video Analytics On Jetson Nano · NVIDIA
Contact
Start a conversation
If you are building something that needs to see, reason, or interact with the physical world, I would like to hear about it.
Irvine, California