// robotics · iiot · edge ml

Manu Vazhunnavar

ME @ Purdue University · Industrial IoT · Robotics

I build sensor systems that turn machine behavior into data. Currently: a non-invasive IIoT monitoring pipeline for industrial compressors — contact-acoustic sensors, DSP feature extraction, and a CNN fault classifier running at the edge. Previously: autonomous platforms, robotic arm control, embedded sensing.

View Projects LinkedIn →

// work

Featured Projects

IIoT · Edge ML · Sensors

IIoT Predictive Maintenance — Atlas Copco G15 FF

Non-invasive fault detection for rotary screw compressors. Maijker contact-acoustic sensors → FFT feature extraction → 1D-CNN classifier → MQTT telemetry. Built for Purdue ME 59700.

PyTorch MQTT Raspberry Pi InfluxDB Grafana DSP

Robotics · Autonomy

Case studies in progress

Autonomous ground platforms, robotic arm control, embedded sensor fusion. Writing up detailed breakdowns — check back soon.

ROS 2 OpenCV C++ URDF

// stack

Technical Skills

Languages

Python C++ C MATLAB JavaScript

ML / Signal

PyTorch NumPy / SciPy FFT / DSP CNN Anomaly Detection

IIoT

MQTT InfluxDB Grafana Node-RED OPC-UA

Robotics

ROS 2 OpenCV Gazebo URDF / SDF

Embedded

Raspberry Pi Arduino I²C / SPI / UART STM32

Infrastructure

Docker Linux Git Node.js SQLite

// about

Background

Mechanical engineering student at Purdue University, enrolled in ME 59700 (Industrial IoT). My work sits at the boundary between physical systems and software — I care about getting real sensor data off real machines and making it useful.

Current focus: predictive maintenance pipelines for industrial equipment, edge-deployed inference, and the signal processing work that makes reliable ML on noisy vibration data actually possible.

Connect on LinkedIn →