Latest
View all →
Biomech Viewer
.png)
Two-stage ML pipeline • R²=0.76
YOLOv8 • mAP50 0.813
Multi-camera triangulation system
3D skeleton viz • FastAPI + Three.js
Pilot briefing app for general aviation
8-camera control system UI
Tablet control system for trainers
Custom website for vintage cake business
Sports biomechanics analysis
3D swing visualization system
Natural language database queries
CAD design & 3D printing
Finite element analysis study
Short videos of recent builds and workflows.
Technical essays, research papers, and public notes on what I am learning and building.
A practical walkthrough of how prompts become tokens, embeddings, vector searches, retrieved context, and grounded model responses. Written for builders who want to understand the system around the model, not just the chat interface.
A practical breakdown of where function calling, parallel agents, and the Model Context Protocol fit in the AI stack, with a focus on access, abstraction, and reusable tool interfaces.
A long-form explanation of skills as reusable instruction packages for coding agents, and how they help keep repeated workflows consistent without bloating the active context window.
This study developed a two-tier machine learning framework to quantify the contribution of forearm muscle activation patterns to elbow varus torque beyond biomechanical kinematics alone. Stage 1 achieved R² = 0.764 using kinematic features, while Stage 2 explained ~30% of residual variance using EMG and isometric performance metrics, supporting a neuromuscular efficiency construct for personalized injury risk assessment.
This study introduces a novel four-bar linkage mechanism to improve adaptive kayak mounts for individuals with upper-body mobility impairments. Using motion capture technology to derive a Standard Stroke Profile, the design replicates natural forward kayak strokes with high accuracy (RMSE = 22.0 mm), addressing key limitations in existing adaptive sports technology.
Pilot briefing app for general aviation. Streamlining pre-flight preparation and enhancing safety. Pursuing FAA approval.
flybebrief.com ↗
Custom streetwear brand. Co-founded with two partners. End-to-end operations from product design to fulfillment.
uncreated.shop ↗
Health-tech startup for real-time concussion detection. Co-founder & CEO. Top 5 Entrepreneur Under 25, Orlando Business Journal.
I build AI, data, and product systems that make hard things easier for people to use.
My work usually starts in messy real-world information and ends as something a coach, teammate, or end user can understand, trust, and act on.
I studied mechanical engineering at Embry-Riddle, played college baseball, and now work as a forward deployed engineer. That background shaped how I think: lead with ownership, build close to the user, move fast, and keep improving the system after the first version ships.
I consider myself an early adopter of LLM systems. A lot of my energy goes into agentic workflows, MCP servers, RAG systems, and LLM-assisted coding. I am interested in how AI can help a single builder move faster without losing technical judgment, taste, or accountability.
I enjoy explaining what I am learning through demos, technical writing, research papers, LinkedIn posts, and Substack. Writing helps me clarify my own thinking, and it makes the work more useful to people who were not in the room while it was built.
My faith is a large part of who I am. I try to approach work with discipline, humility, and consistency. Titles change. What matters more is becoming the kind of person who can be trusted with the work in front of him.