Autonomous Cross-Species Taskmaster Built with Gemini 3.7 Flash

Over 70% of companion dogs suffer from behavioral anxiety, territorial reactivity, or separation distress. But there is a glaring design limitation in modern software:

Pets can’t type into a chatbot or navigate a smartphone menu when they are in distress.

When a delivery driver rings the doorbell at 92 dB or a sudden thunderstorm rolls in, the critical window to de-escalate canine sympathetic nervous system arousal is measured in seconds. If an owner is away or on a video call, cortisol levels spike, reinforcing learned reactivity.

For the All Things Agentic Hackathon**, we built PetWhisperer AI — an autonomous, hands-free Agentic Taskmaster that passively listens and watches for environmental triggers, diagnoses emotional distress using Google Gemini 3.7 Flash, and coordinates a 5-stage remediation pipeline in real-time.

---

🏗️ System Architecture Overview

PetWhisperer operates on an event-driven loop that bridges sensory ingestion, cognitive reasoning, and physical bio-acoustic intervention:

The 5-Stage Autonomous Execution Pipeline:

1. Sensory & IoT Ingestion: Passive acoustic FFT hydrophone monitors decibel thresholds (e.g., a 92 dB acute spike).

Cognitive Ethology Diagnosis*: Gemini 3.7 Flash* calculates an Arousal Index ($0-100$) and Cortisol Risk.

1. Bio-Acoustic Intervention**: Native **Web Audio API synthesizes restorative 432 Hz Solfeggio harmonic sine tones.
2. Data Warehouse Telemetry: Structured event vectors are streamed to Snowflake** for population-level behavioral modeling.
3. On-Chain Behavioral Verification: An ed25519 signature anchors the event to Solana Devnet** and awards $TREATS tokens.

---

🧠 Leveraging Gemini 3.7 Flash & 2.5 Flash

We combined gemini-3.7-flash for cognitive task orchestration with **gemini-2.5-flash** for sub-second vision processing via the official@google/genai` TypeScript SDK:

`typescript
import { GoogleGenAI } from '@google/genai';

const ai = new GoogleGenAI({ apiKey: process.env.GEMINI_API_KEY });

export async function triageAutonomousIncident(triggerType: string, intensity: number, dogProfile: any) {
const prompt = `
You are an expert veterinary ethologist and autonomous coordinator.
Evaluate the following incident:
- Dog: ${dogProfile.name} (${dogProfile.breed}, Age ${dogProfile.ageYears})
- Trigger: ${triggerType}
- Intensity: ${intensity}%

Output structured JSON conforming to the ethology schema.

`;

const response = await ai.models.generateContent({
model: 'gemini-3.7-flash',
contents: prompt,
config: {
responseMimeType: 'application/json',
temperature: 0.2
}
});

return JSON.parse(response.text || '{}');
}