SiliconWit.COM SiliconWit.COM
John Kibe's avatar

John Kibe

Mechatronic Engineer & Industrial Automation Specialist

Mechatronic engineer specializing in industrial automation, electronic design, and motor control systems.

About John Kibe

John Kibe is a mechatronic engineer with expertise in industrial automation with PLCs, electronic design, motor control, and fluid power operations. He holds a B.Sc. in Mechatronic Engineering from Jomo Kenyatta University of Agriculture and Technology (JKUAT, 2014) and is a Siemens Certified Mechatronic Systems Associate from Dedan Kimathi University of Technology (2018). A registered Graduate Engineer with the Engineers Board of Kenya (EBK, 2015), John brings hands-on experience bridging mechanical, electrical, and software engineering in industrial settings.

Focus

Industrial Automation with PLCs
🎨 Electronic Design
Motor Control
Fluid Power Operations
📡 IoT Systems & MQTT

📚 Education Contributions

Iot Systems

Alerts, Automation, and Rule Engines

Configure threshold alerts that notify via email, SMS, Discord, Slack, and Telegram. Build automation flows with Node-RED that trigger actions based on sensor data. Compare self-hosted rule engines with the SiliconWit.io alert system.

Published: March 10, 2026

iotalertsautomationnode-redmqttnotificationsrule-enginegrafanasiliconwit-io

Device Security, TLS, and Provisioning

Generate and deploy X.509 certificates for mutual TLS authentication. Implement device identity and provisioning workflows. Secure firmware updates with signed images. Audit your IoT deployment against common attack vectors.

Published: March 10, 2026

iotsecuritytlsmtlsx509certificatesprovisioningesp32mosquittoowasp

IoT Architecture and Protocol Comparison

Survey IoT system architectures and compare MQTT, CoAP, and HTTP side by side. Send the same BME280 sensor data over all three protocols from an ESP32 and measure bandwidth, latency, and power consumption to make informed protocol choices.

Published: March 10, 2026

iotmqttcoaphttpesp32bme280protocolsarchitecturesiliconwit-io

MQTT Broker Setup and Secure Connections

Install and configure a Mosquitto MQTT broker with TLS encryption, password authentication, and topic ACLs. Connect to both a self-hosted broker and the SiliconWit.io platform. Understand retained messages, last will, QoS levels, and persistence.

Published: March 10, 2026

mqttmosquittotlsiotbrokersecurityacldockersiliconwit-io

MQTT Clients on ESP32, Pico, and STM32

Write MQTT client firmware for three MCU platforms: ESP32 with ESP-IDF, RPi Pico W with MicroPython, and STM32 with an ESP-01 Wi-Fi module. Design a consistent topic hierarchy, publish structured JSON payloads, handle reconnection with exponential backoff, and buffer messages when offline.

Published: March 10, 2026

iotmqttesp32rpi-picostm32multi-platformjsonqosreconnection

Capstone: Production IoT Monitoring System

Combine everything from the IoT Systems course into a production deployment: multiple sensor nodes, TLS-secured MQTT broker, time-series database, Grafana dashboards, automated alerts, a REST API, and cloud forwarding to SiliconWit.io.

Published: March 10, 2026

iotmqttproductiongrafanainfluxdbdockercapstoneesp32raspberry-pisiliconwit-io

Real-Time Dashboards and Data Visualization

Store MQTT sensor data in InfluxDB via Telegraf, build Grafana dashboards with live charts and gauges, explore the SiliconWit.io dashboard as a managed alternative, query historical data with Flux, and create lightweight Chart.js dashboards for embedded gateways.

Published: March 10, 2026

iotgrafanainfluxdbtelegrafdashboarddata-visualizationmqtttime-seriesfluxchart-js

REST APIs, Webhooks, and Device Integration

Design RESTful APIs for IoT device management and data retrieval. Build webhook endpoints that receive push notifications, validate HMAC signatures, and integrate REST with MQTT for a complete IoT data pipeline.

Published: March 10, 2026

iotrest-apiwebhooksflaskhmacesp32device-managementsiliconwit-io
Edge Ai Tinyml

Edge-Cloud Hybrid Architectures

Cover tiered inference, cloud-assisted retraining, OTA model updates, and federated learning concepts. Build a complete system where an ESP32 runs local anomaly detection, escalates uncertain cases to a cloud server for a more powerful model, receives retrained models via OTA, and reports status to an MQTT dashboard.

Published: March 9, 2026

tinymledge-aiedge-cloudhybridotafederated-learningmqttesp32model-update

Camera Image Classification on ESP32

Deploy a MobileNet-based image classifier on an ESP32-CAM module with OV2640 camera and PSRAM. Capture images, preprocess them on device, run TFLite Micro inference, and classify objects or detect people in real time within the ESP32's tight memory budget.

Published: March 8, 2026

tinymledge-aiimage-classificationesp32-cammobilenetcnntensorflow-litecomputer-visionperson-detectionpsram

Anomaly Detection for Predictive Maintenance

Train an autoencoder on normal vibration data from an MPU6050 mounted on a motor, deploy it on an ESP32, and detect mechanical anomalies in real time. Compare edge inference with cloud-based anomaly scoring for latency, bandwidth, and reliability.

Published: March 7, 2026

tinymledge-aianomaly-detectionpredictive-maintenanceautoencoderesp32mpu6050vibrationmqttindustrial

Accelerometer Gesture Recognition

Collect labeled gesture data from an MPU6050 accelerometer, train a TensorFlow classifier, and deploy it on both RPi Pico and STM32 using TensorFlow Lite Micro. Compare inference performance across platforms with LED feedback and MQTT publishing.

Published: March 6, 2026

tinymledge-aigesture-recognitionmpu6050rpi-picostm32accelerometertensorflow-litemqtt

Keyword Spotting and Voice Wake Word

Build a Hey Device wake word detector using an ESP32 and INMP441 I2S MEMS microphone. Capture audio with the I2S driver, extract MFCC features on device, train a keyword model on the Speech Commands dataset, and run real-time inference to trigger actions on detection.

Published: March 5, 2026

tinymledge-aikeyword-spottingwake-wordesp32i2smfccspeech-commandstensorflow-litemicrophone

Model Quantization and Optimization for MCUs

Apply post-training quantization and quantization-aware training to a CNN model. Compare float32 vs int8 accuracy, inference speed, flash size, and RAM usage on ESP32. Learn model pruning basics and best practices for deploying optimized models on microcontrollers.

Published: March 4, 2026

quantizationtinymlint8post-training-quantizationquantization-aware-trainingmodel-optimizationesp32edge-ai

TensorFlow Lite Micro Model Deployment

Deep dive into the TFLite Micro runtime architecture. Train a gesture classifier in TensorFlow, convert it to TFLite Micro, and deploy it on both ESP32 and STM32. Compare inference time, RAM usage, and porting considerations across platforms.

Published: March 3, 2026

tflite-microtensorflow-liteesp32stm32gesture-recognitiontinymledge-aicross-platform

Edge Impulse Data Collection and Training

Collect accelerometer data from an MPU6050 on ESP32, upload it to Edge Impulse, train a motion classifier for idle, walking, and running activities, and deploy the quantized model back to the ESP32 for real-time inference.

Published: March 2, 2026

edge-impulsetinymlesp32mpu6050accelerometermotion-classificationedge-ai

TinyML and Machine Learning on Microcontrollers

Understand the TinyML landscape, hardware constraints, and the full ML pipeline from TensorFlow training to on-device inference. Deploy a sine wave regression model on an ESP32 using TensorFlow Lite for Microcontrollers.

Published: March 1, 2026

tinymledge-aitensorflow-liteesp32machine-learningmicrocontrollerstflite-micro

Contribution Stats

34
Authorship SSU
2025
Member Since
Kenya
Location
17
Education Contributions

Connect & Follow