IoT Developer for Predictive Maintenance Gateway
Budget: $250 – $750 SGD
Senior IoT Programmer/Engineer – Edge ML & Modbus TCP Gateway (Edge Impulse, Grafana with InfluxDB)
Summary
We are seeking an expert IoT Developer to build a production-grade, air-gapped Predictive Maintenance gateway. The system must collect high-frequency vibration data from 19 industrial wireless sensors via Modbus TCP, process it using a local Edge AI model, and manage the fleet via a self-hosted ThingsBoard CE server.
Strict Constraint: No cloud-native dependencies (AWS/Azure/Google) are permitted for data storage or processing. All data must reside on local infrastructure.
Key Deliverables:
1) Gateway Configuration Portal (Custom Web UI): Build a local web-based interface (Flask/FastAPI) hosted on the Raspberry Pi. This portal must allow a non-technical user to configure Sensor IPs, Modbus Registers, and ThingsBoard Credentials via a browser—no CLI/Linux terminal access required.
2) Modbus & ML Engine: Implement a high-performance asynchronous Modbus TCP master and integrate an Edge Impulse C++ SDK for local anomaly detection.
3) Local Data Orchestration: Deploy and configure a local ThingsBoard CE instance (via Docker) to act as the central on-premise management hub.
4) Industrial Logging (InfluxDB): Set up a local InfluxDB instance with a defined Retention Policy (Raw data for 7 days, aggregated for 90) to prevent SSD overflow.
5) Hardware Optimization: Configure the system for 24/7 reliability on a Raspberry Pi 5 with an NVMe SSD, including a hardware watchdog and RAM-buffered writing.
Technical Requirements:
Protocols: Industrial Modbus TCP/IP, MQTT.
Platforms: ThingsBoard CE (Self-hosted), Edge Impulse, Docker-Compose.
Languages: Python (AsyncIO) or C++ for the edge runtime.
Database: InfluxDB with time-series optimization.
Summary
We are seeking an expert IoT Developer to build a production-grade, air-gapped Predictive Maintenance gateway. The system must collect high-frequency vibration data from 19 industrial wireless sensors via Modbus TCP, process it using a local Edge AI model, and manage the fleet via a self-hosted ThingsBoard CE server.
Strict Constraint: No cloud-native dependencies (AWS/Azure/Google) are permitted for data storage or processing. All data must reside on local infrastructure.
Key Deliverables:
1) Gateway Configuration Portal (Custom Web UI): Build a local web-based interface (Flask/FastAPI) hosted on the Raspberry Pi. This portal must allow a non-technical user to configure Sensor IPs, Modbus Registers, and ThingsBoard Credentials via a browser—no CLI/Linux terminal access required.
2) Modbus & ML Engine: Implement a high-performance asynchronous Modbus TCP master and integrate an Edge Impulse C++ SDK for local anomaly detection.
3) Local Data Orchestration: Deploy and configure a local ThingsBoard CE instance (via Docker) to act as the central on-premise management hub.
4) Industrial Logging (InfluxDB): Set up a local InfluxDB instance with a defined Retention Policy (Raw data for 7 days, aggregated for 90) to prevent SSD overflow.
5) Hardware Optimization: Configure the system for 24/7 reliability on a Raspberry Pi 5 with an NVMe SSD, including a hardware watchdog and RAM-buffered writing.
Technical Requirements:
Protocols: Industrial Modbus TCP/IP, MQTT.
Platforms: ThingsBoard CE (Self-hosted), Edge Impulse, Docker-Compose.
Languages: Python (AsyncIO) or C++ for the edge runtime.
Database: InfluxDB with time-series optimization.
Related categories:
Python
C++ Programming
Raspberry Pi
Docker
Flask
Anomaly Detection
Data Management
FastAPI