Biomedical Engineering / UT Austin

Building movement intelligence systems.

Biomedical Engineering @ UT Austin. Working across biomechanics, wearable sensing, AI systems, and real-world decision intelligence.

Research

Signals, control, and sensing.

Research around biomechanics, sensing, and transferable structure in human movement.

Representation Learning / Motor Control

Perturbation Recovery CVAE

Learning whether early neuromuscular signals encode transferable control policies across humans.

Details

Built a CVAE using early EMG signals from 0-80ms post-perturbation on the Lorenz treadmill perturbation dataset. The latent representation predicts cross-subject muscle synergy activations while direct mechanics prediction fails.

Why it matters

Suggests control structure can transfer across humans while mechanical execution remains body-specific.

CVAEEMGMotor ControlRepresentation LearningBiomechanics
In progress

Wearable Biomechanics

Lower-Limb Loading

Building longitudinal baselines from wearable insole data to detect drift in athlete loading patterns.

Details

Focuses on loading rate, impulse, stance time, asymmetry, fatigue-related drift, and field monitoring across drills and surfaces.

Why it matters

Moves athlete monitoring from one-off testing toward personalized field baselines.

LoadsolGRFWearablesDrift DetectionAthlete Monitoring
Active

Biomedical Sensing / Hardware

Graphene Biosensors

Working on graphene biosensor fabrication, PCB design, and sensor programming in Akinwande Lab at UT Austin.

Details

Connects hardware-aware biomedical sensing with physiological signal acquisition and interpretation.

Why it matters

Grounds sensing and AI work in the realities of signal acquisition hardware.

GraphenePCB DesignBiosensorsSignal AcquisitionHardware
Active

Projects

Decision systems and AI infrastructure.

Products and prototypes shaped by the same thesis: turn context and movement signals into decisions.

Venture / Movement Intelligence

Kynetic / Borelli Labs

Movement intelligence platform for athletic trainers, coaches, and rehab teams.

Problem

Athletic trainers manage too many athletes with too little time and too much raw data.

Solution

Turns wearable and biomechanical data into decision-grade outputs: modified practice, load caps, hold out, or medical review.

Why it matters

The main venture thesis: movement data should produce decisions, not dashboards.

SaaSFastAPIPostgresReactWearablesBiomechanics
Active venture

AI Infrastructure

Atlas

Universal memory layer for AI tools.

Problem

Every AI session starts from zero; context is fragmented across Slack, Notion, ChatGPT, GitHub, and local work.

Solution

Captures, stores, compresses, and injects relevant context into AI sessions using profile-isolated filesystem memory, semantic scoring, and temporal decay.

Why it matters

Explores context as infrastructure rather than prompt-by-prompt memory.

AI InfrastructureContext EngineeringMCPBrowser ExtensionSemantic Search
Prototype

Applied AI / Hackathons

AI Agent Systems

Rapid prototyping of multimodal and live LLM agent systems.

Problem

Most AI demos are static; real workflows require streaming, memory, tool use, and interaction.

Solution

Built Gemini/live-agent and hackathon systems for real-time AI interfaces.

Why it matters

Tests live interfaces where LLMs listen, reason, and act inside real workflows.

TypeScriptLLMsGeminiAgentsStreaming
Prototype

Stack

Technical range.

A compact toolset across biomechanics, sensing, AI systems, and product engineering.

AI / ML

PyTorch / CVAE / LightGBM / LLM agents / Signal modeling

Biomechanics

EMG / GRF / Loading rate / Wearables / Stride data

Engineering

Python / TypeScript / React / FastAPI / Postgres / Docker

Hardware

PCB design / Graphene sensors / Signal acquisition / IMUs