Western Cyber Society

AI Engineer · September 2025 — Present

AI Engineer – Motorshield

Western Cyber Society

Architecting an end-to-end predictive maintenance platform for motor telemetry, processing high-frequency time-series sensor data to detect faults before they cause failures.

Key Responsibilities

• Data Pipeline: Architected a scalable data pipeline (Simulink → ingestion services → TimescaleDB → ML models) to process high-frequency time-series sensor data (voltage, torque, RPM, temperature).
• Data Preprocessing: Built a preprocessing pipeline (windowing, labeling, augmentation) converting raw signals into model-ready datasets with overlapping 40ms windows and fault classification labels.
• Deep Learning Models: Developed and evaluated deep learning models — CNN, Transformer, and Hybrid CNN–Transformer (~464k params) — for multiclass fault detection using 6-channel time-series inputs.
• Data Augmentation: Implemented robust data augmentation strategies (noise injection, harmonics, drift, transient spikes) to improve real-world generalization across hardware variability.
• Award: 🏆 Industry Choice Award, Canadian Tech Summit — selected by industry judges for exceptional innovation and real-world impact beyond standard evaluation criteria.