"use client"; import Link from "next/link"; import { Cpu, Zap, Activity, ShieldCheck, ArrowRight, Terminal, GitBranch, Sliders } from "lucide-react"; export default function Home() { return (
{/* Hero Section */}
{/* Eyebrow Badge */}
Edge AI Clinical Intelligence / ESP32 & ARM Cortex-M

CLINICAL
INFERENCE
AT THE EDGE.

Cardiovascular anomaly classification transpiled directly for bare-metal microcontrollers. Stream multi-channel physiological vitals, evaluate a calibrated 5-model soft-voting ensemble, and export zero-malloc C99 headers for embedded silicon.

{/* Technical Telemetry Badges Matrix */}
INFERENCE LATENCY
18.4 μs
ESP32-S3 @ 240MHz
STATIC RAM
28 Bytes
Zero Dynamic Heap
SERIAL BUS
115.2k
Web Serial API
ECG SAMPLING
250 Hz
Lead-II Waveform
{/* Action Button Grid */}
Launch Monitor Simulator C99 Compiler MLOps Registry
{/* Feature Grid: Three Structural Modules */}
{/* Card 01 */}
01 · QUANTIZATION ENGINE

INT8 Quantized Weights

Continuous weight vectors and decision thresholds quantized to 8-bit integers. Delivers 75% SRAM reduction for constrained microcontrollers without floating-point FPU reliance.

MEMORY SCALE: 4:1 INT8 OPEN COMPILER →
{/* Card 02 */}
02 · ENSEMBLE ARBITRATION

5-Model Soft-Voting

Aggregates class probabilities across Random Forest (30 trees), Linear SVM, Logistic Regression, MLP Neural Network, and KNN to suppress single-estimator variance.

TOPOLOGY: 5 ESTIMATORS OPEN SIMULATOR →
{/* Card 03 */}
03 · CODE GENERATION

Zero-Malloc C Headers

Transpiles scikit-learn ensembles directly to standalone C99 headers (model.h) with zero dynamic heap allocation, avoiding fragmentation crashes on bare-metal firmware.

TARGET: ESP32 // CORTEX-M VIEW C99 EXPORTER →
{/* Understated Architecture Strip */}
CORE ARCHITECTURE:
UART 115,200 (8-N-1) LEAD-II SAMPLING (250 HZ) 5-CHANNEL VITAL SIGN VECTOR ZERO-MALLOC C99 ABI SHAP TREE ATTRIBUTION
); }