Mohammad Hamzah Warsi

Electronics Engineer | Hardware Design, Embedded Firmware, Test & Automation

Profile summary

Career highlights

Key skills

Professional experience

Managed automation tasks for live TV production line systems, overseeing hardware troubleshooting and data organization to support daily operational needs.

• •Cut production downtime by 30% in 3 months by deploying and fine-tuning Fanuc and Mitsubishi robots on a live TV production line.
• •Kept production running by leading control panel and power system troubleshooting, and supported electrical cabinet and control panel work for the robot cells.
• •Sped up root-cause analysis by designing an Excel dashboard that tracked every robot fault and stoppage.

Executed quantitative modeling for electricity prices and water availability to support energy planning for hydropower and wind integration. Built technical foundations using Python, scikit-learn, pandas, and NumPy for forecasting systems.

• •Delivered higher forecast precision by owning the end-to-end data pipeline, covering data cleaning, preprocessing, feature selection, and model tuning.
• •Established machine learning workflows with data versioning and experiment tracking, ensuring reproducible and comparable results.

Developed firmware and conducted hardware testing for integrated cell tester towers, supporting deployments for automotive and industrial clients. Programmed embedded systems on processors, validated safety protocols, and resolved electrical and thermal control issues across test stands.

• •Programmed C firmware across embedded Linux and RTOS on TI AM64x processors to develop docking controller state machines and reduce test cycle variability.
• •Calibrated AI8 analog input modules in Keil uVision and refined lookup tables to improve input accuracy.
• •Implemented a vacuum system monitoring function supporting pressure ranges from -1 to 12 bar with a precision of 1 percent.

Projects

Designed a multi-source system (piezoelectric, thermoelectric, wind) that captures waste heat, vibration and airflow to power street lighting and utilities, projected at 350W at highway speeds. Engineered the power conditioning and 12V battery charging circuit, including a boost converter that holds a 10–20V input at 24V with under 2 mV ripple, and modelled the system in ANSYS (mechanical and thermal), LTspice and Simulink.

Trained five deep learning models (PentaCNN, MobileNetV2, VGG-16, VGG-19, DenseNet121) in Python on the JmuBen dataset using GPU and TPU, deployed the best model to GPU, TPU and VPU, and built a benchmarking solution comparing all three platforms. With TensorFlow Lite and 8-bit quantisation, the Coral Edge TPU reached 10 ms inference against 40 ms on a GPU workstation, and MobileNetV2 reached 98.9% training accuracy (first-author IEEE paper).

Engineered a 12V/2A supply (TL494 controller, 100 kHz) incorporating a transformer, supercapacitor bank, inverters and relay control coils, with analytical calculations of component ratings, efficiency and thermal performance. Designed and routed a 4-layer PCB in KiCad, and reached ~85% efficiency with 150 mV ripple.

Trained a custom detection model from YOLOv3 weights on a large self-built dataset of face images, augmented for lighting, angle and expression to improve robustness. Implemented shape detection for circles, squares and triangles, with real-time detection in OpenCV and Python.

Built a speed gun from a Doppler radar module (HB100) and an STM32 microcontroller, with op-amp signal conditioning and a signal processing algorithm that converts the Doppler frequency shift to velocity using zero-crossing detection. Measured speed within ±5% of a commercial radar gun and displayed it in real time on an LED interface.

Designed and etched copper PCBs for an 8-sensor infrared array, wrote Arduino firmware for sensor processing and PID motor control, and modelled and 3D-printed a custom IR sensor holder in AutoCAD, achieving stable line tracking on the test course.

Education

Electrical & Electronic Engineering

Certifications

Languages