🏆 Google Gemini 3 Hackathon · 2025
Neuro-Sync Live
Bio-Adaptive AI Audio Entrainment · Progressive Web App
🔒 Patent Pending · UK
"What if your music player could see how you feel and adapt your brainwaves in real-time?"
Existing focus audio tools play static loops whether you are exhausted, overstimulated, or in a deep flow state. Neuro-Sync Live introduces an active closed feedback loop: a privacy-first AI system that reads real-time bio-feedback via camera feed and dynamically calculates and shifts binaural audio frequencies (Alpha 10Hz, Theta 6Hz, Gamma 40Hz) to regulate cognitive arousal.
☁️ Cloud Multimodal Mode
Gemini 1.5 Flash streaming via Vercel Edge Functions for deep visual affect and posture analysis.
🔒 Zero-Latency Local Mode
Gemma 2B running offline inside the browser on WebGPU — completely airgapped with zero external server calls.
🎵 Custom Audio Synthesizer
Native Web Audio API oscillators producing precision micro-frequency differential tones in real time.
🧠 Dynamic Entrainment
Rapidly transitions between focus, meditation, and alertness within seconds of detected physiological changes.
Next.js 14
Gemini 1.5 Flash
Gemma 2B (WebGPU)
MediaPipe
Web Audio API
Tailwind CSS
TypeScript
🛡️ Applied AI Safety · IP
PromptGuardian
LLM Adversarial Prompt Injection Defense Framework
🔒
"Securing autonomous agents before malicious inputs hijack execution."
A production-engineered AI security tool designed to detect, classify, and neutralize indirect and direct prompt injection attacks. Built to protect enterprise autonomous agent pipelines from prompt manipulation, system prompt leaks, and unauthorized tool execution.
Python
LLM Safety & Red-Teaming
Adversarial Evaluation
DeepEval
Circuit-Breakers
