全端工程師

Role & Scope

• Own everything above the Edge AI runtime: cloud backend, web dashboards, public-facing venue displays, smartwatch apps, and the incident workflow engine
• Connect AI detection to human response, resolution, and audit — the full incident closure loop
• Not supporting an existing team — building the platform that turns edge AI output into an operational product from scratch
• Work directly with CEO and CTO to define what to build, then go build it

First 6 Months — Deliverables

Q1:

• User Dashboard — Multi-site event management portal with traffic trend analytics, alert accuracy metrics, and evidence package review (snapshots, video, timeline, zone overlay)
• Venue Display — Public-facing digital signage for facility lobbies showing real-time AI monitoring status, occupancy data, and promotional content rotation. Signage design language, not an ops console

Q2:

• SmartWatch App — Cross-platform wearable app for alert acknowledgment, dispatch, and case closure. One-handed operation, under 2 seconds per critical action
• Emergency Workflow Engine — Bidirectional event lifecycle engine implementing an 8-state Case Workflow State Machine (Queued → Needs Confirm → Validated → Rejected → Dispatched → Escalated → Closed → Audit Pending) with SLA timers, timeout escalation, and dispatch routing

Backend

• Build Event Ingest API receiving structured JSON and evidence payloads from edge devices via IPSec/WireGuard VPN
• Implement GCP Pub/Sub event streaming with dual storage: BigQuery for time-series analytics, Postgres for operational state
• Develop Workflow State Machine covering the full 8-state case lifecycle with configurable SLA timers, timeout escalation triggers, and dispatch routing logic
• Build multi-tenant management layer: sites, users, roles, and permissions
• Implement notification hub spanning LINE, email, webhook, and smartwatch push
• Expose RESTful and WebSocket APIs serving all three client surfaces (dashboard, display, watch)

Frontend

• Web Dashboard — Real-time event feeds, historical analytics charts, event detail views with evidence packages, and a CMS module for Venue Display content management
• Venue Display — Large-screen optimized rendering with automatic offline fallback and local caching when cloud connectivity drops
• SmartWatch App — Cross-platform compatibility targeting Wear OS native or web-based approach; technology choice is yours to make and defend

Infrastructure

• Manage GCP project end-to-end: Cloud Run, Pub/Sub, BigQuery, Cloud SQL, Cloud Storage
• Build and maintain CI/CD pipelines for all cloud services
• Implement fleet-wide monitoring, alerting, and health dashboards across all deployed sites
• Handle OTA content delivery for Venue Display

Tech Stack

• Edge (interface with, not own) — NVIDIA Jetson Orin Nano, DeepStream, YOLO + VLM pipeline. Pushes structured event JSON + evidence packages via IPSec/WireGuard VPN
• Cloud / Backend — GCP (Pub/Sub, Cloud SQL/Postgres, BigQuery, Cloud Run). Python or Node.js. WebSocket for real-time streaming
• Frontend — React or Vue (pick one and commit). Data visualization library. Responsive dashboard + signage-optimized display
• Wearable — Wear OS native or cross-platform. BLE/WiFi connectivity within facility networks

Requirements — Must Have

• 3+ years shipping production web applications end-to-end (backend + frontend + deployment)
• Track record of building cloud backends from scratch, not just adding features to existing systems
• Hands-on experience with at least one major cloud platform: GCP, AWS, or Azure (we use GCP)
• Real-time systems experience: WebSocket, event-driven architecture, Pub/Sub or message queues
• Relational database design and query optimization (Postgres preferred)
• Proficiency in both Python and JavaScript/TypeScript
• At least one shipped mobile or wearable application in production

Requirements — Strongly Preferred

• Workflow or state machine system development (ticket systems, dispatch engines, approval flows — anything with lifecycle states and timeout logic)
• Video/media handling: HLS streaming, thumbnail generation, evidence packaging
• Prior experience owning a full product surface at a startup with fewer than 10 people
• GCP-specific hands-on: Pub/Sub, Cloud Run, BigQuery
• Mandarin fluency for internal communication + English technical reading proficiency

Requirements — Nice to Have

• IoT or edge integration experience (MQTT, device management, OTA)
• Digital signage or kiosk application development
• Background in safety/compliance systems, incident management, or operational workflows