AI Developer, Full-Stack Engineer, or Agency

Overview

We are seeking an expert AI Developer, Full-Stack Engineer, or Agency to build a custom, on-premise multimodal monitoring system for our manufacturing plant. The system must process live camera feeds locally, display real-time events on a central web dashboard, and route automated, time-based safety and operational, site monitoring, alerts through WhatsApp using a multi-level manager escalation engine.

This contract is split into two distinct phases:

• Phase 1: A one-off, milestone-based fixed contract to develop and fully deploy the system.
• Phase 2: A long-term monthly maintenance retainer ($400 - $800/month) to manage system uptime, monitor APIs, and support the plant team.

Key Core Deliverables (Phase 1)

1. Real-Time Object & People Tracking

   • Multi-camera tracking (using YOLOv8/v10 and ByteTrack/DeepSORT) to follow workers, forklifts, and assets seamlessly across the factory floor.
   • Automated PPE compliance tracking (hard hats, safety vests, boots).
   • Monitoring staff behavior and tracking them in the plant on duties, and loitering or talking in groups.
   • Monitoring Objects and the movement of stock and tracking machines if operational or not.

2. Central Web Dashboard

   • A responsive, live-updating user interface (React.js, Vue, or Python-based Streamlit/Dash) installed on our plant network.
   • Features must include real-time incident event logs, worker density heatmaps, and active alert countdowns.

3. Two-Way WhatsApp Automation & Tracking

   • Seamless integration with Twilio or the native Meta WhatsApp Business API.
   • The system must send automated questions when anomalies occur (e.g., "Machine 3 is idle. Why is this task not done?").
   • The database must actively track worker responses, logs, and answer timestamps.

4. Time-Based Manager Escalation Engine

   • If an operator fails to reply to a WhatsApp query within set minutes, the system must update the live dashboard status and instantly escalate the alert to the Floor Manager via WhatsApp.
   • If the manager fails to respond within another 15 minutes, the system must escalate the issue up the chain to the Plant Director and flag the dashboard as "CRITICAL UNRESOLVED".

Required Technical Expertise

• Deep experience with Python, OpenCV, PyTorch, or TensorFlow.
• Proven mastery of Multi-Object Tracking (MOT) frameworks and real-time streaming protocols (RTSP/WebRTC).
• Strong backend experience building state machines and relational databases (PostgreSQL/Redis) to track pending communication states.
• Prior experience managing Twilio or Meta WhatsApp Business API integrations.

Application Requirements (To filter out automated bot proposals)

Please start your application message with the words "FACTORY MONITOR AI" so we know you read this full post.

In your application, please answer these three specific questions:

1. What tracking algorithm (e.g., DeepSORT, ByteTrack) do you recommend for multi-camera worker tracking without duplicate IDs, and why?
2. How will you structure the database logic to check if a specific employee has failed to reply to a WhatsApp message within a 15-minute window?
3. What is your proposed fixed budget for Phase 1, and your target monthly retainer for Phase 2?