EE / FIRMWARE / INDUSTRIAL AUTOMATION / RENEWABLES

Mohammad Hamzah Warsi — building the low-level systems that keep high-stakes hardware honest.

Electrical & Electronics Engineer · Firmware · Battery Test Infrastructure · Automation

MEng in Electrical & Electronics Engineering (Nottingham). Direct engineering experience across BMW, Sony, Fraunhofer, Varta and PEC NV — shipping bare-metal and Linux firmware, industrial robotics, and validated test infrastructure in environments where a wiring fault means a safety incident, not a bug ticket.

€100M
BMW customer handover
30%
downtime reduction, Sony
84,000 Ah
cell pallet prototype, 20 cells
2
peer-reviewed publications

Experience

01 / TIMELINE

Electronics Engineer — Firmware & Testing

PEC NV, Belgium
Feb 2025 – Dec 2025
  • Developed bare-metal and Linux-based firmware for Integrated Cell Testers (TI MCUs, C) for BMW and Fraunhofer — new docking-controller functions and optimized low-level control logic to reduce test-cycle variability.
  • Integrated pressure transducers (±1% accuracy), vacuum systems, safety systems and HMI monitoring for real-time cell control; built 3-way valve and fan control logic for thermal regulation.
  • Diagnosed and corrected an overheating safety issue (incorrect wiring) on Varta testers; delivered a probe comparison study for Fraunhofer.
BMW Project — Wackersdorf, Germany
  • Prototyped a new cell pallet for 20 cells at an 84,000 Ah charging rate (4.2V max) for BMW's "Neue Klasse" vehicles; refined testing parameters to improve safety margins without compromising throughput.
  • Delivered a 1-week rapid intervention resolving wiring faults, overheat risk, sensor placement and a condensation root cause; completed tower RD01 validation, contributing to the €100M customer handover.
Integrated Cell Tester Towers — Firmware Detail
  • AI8 modules: firmware calibration update (Keil uVision), refined lookup tables and gain offsets, validated with documented before/after accuracy.
  • Safety systems: reformatted alarms for service engineers; added bimetal-flicker and GTS air-pump failure detection; built a DO/DI cross-validation health check and evaporator post-run cooling logic.
  • THC cards: calibrated and validated 40 thermal control cards; found and fixed register misconfiguration and offset errors; automated calibration scripts and reduced manual steps.
  • Pressure transmitters: real-time monitoring across –1 to 12 bar, ±1% accuracy, live HMI integration with overpressure alerts.
  • Type C pallet: contributed to the next-generation pallet deployed at BMW, reducing contact-resistance variability and authoring best-practice test documentation.

Automation Engineer Intern

Sony EMCS, Malaysia
Jun 2022 – Sep 2022
  • Deployed and optimized Fanuc and Mitsubishi robots on a live TV production line, achieving a 30% reduction in downtime within 3 months.
  • Led control panel and power system troubleshooting to keep production uninterrupted.
  • Designed an Excel-based error dashboard enabling faster root-cause analysis.

Machine Learning Engineer

Think Outside, Norway
Feb 2022 – May 2022
  • Built time-series forecasting models for electricity prices and water availability, supporting 5-year energy planning for hydropower and wind integration.
  • Owned the end-to-end data pipeline — cleaning, preprocessing, feature selection, model tuning — improving forecast accuracy.
  • Established reproducible ML workflows with data versioning and experiment tracking.

Featured Projects

02 / APPLIED WORK
Power · Embedded · Simulation

Energy Harnessing System for Smart Highways

Multi-source renewable energy system for Malaysian highways, harvesting wasted vehicle energy to power street utilities. Piezoelectric transducers with a full-wave bridge rectifier converted pulsed mechanical input (~65V) to a regulated 12V output, while a closed-loop DC-DC boost converter stabilized a wind turbine's fluctuating 10–20V output to a constant 24V with under 2mV ripple. Wind turbine simulations (Ansys Fluent) projected up to 350W at highway speeds from a NACA 4412 horizontal-axis design. All sub-systems — piezoelectric, thermoelectric, wind — were integrated onto a single PCB and validated in LTspice, Simulink, and partial physical prototyping.

350W projected output<2mV rippleLTspice / Simulink / Ansys Fluent
Power Electronics

Switch-Mode Power Supply

12V/2A buck converter using a TL494 PWM controller and MOSFET switching at 100kHz, with a custom 4-layer PCB routed in KiCAD. Reduced gate-drive parasitic inductance and added a freewheeling diode after oscilloscope-based waveform troubleshooting.

~85% efficiency150mV rippleKiCAD
Embedded · ML

Edge Coffee Leaf Disease Classification

Trained five deep learning models (MobileNetV2, VGG-16/19, DenseNet121, PentaCNN) on the JmuBen dataset, then deployed and benchmarked the best performer across GPU, TPU and VPU using a Google Coral Developer Board Mini for real-time, on-device inference.

98.9% training accuracy10ms vs 40ms inference
Signal Processing · STM32

Doppler Radar Speed Gun

Continuous-wave Doppler radar speed gun using an HB100 microwave sensor (10.525GHz) and an STM32. Designed an op-amp band-pass signal-conditioning stage and a zero-crossing algorithm to convert Doppler shift to velocity, calibrated against a known moving target.

±5% error vs commercial gun
Robotics · Controls

Hercules 4WD Line-Following Robot

Autonomous 4WD robot on an Arduino Mega with a PID control loop over an 8-sensor IR array, driving four DC geared motors through an L298N driver. Self-etched PCBs for the sensor array and a 3D-printed AutoCAD sensor mount.

PID controlCustom PCB + mount

Publications

03 / RESEARCH
Published

Advances in Energy Harnessing Techniques for Smart Highways: A Review

Springer Nature: Electrical Engineering · DOI 10.1007/S00202-024-02379-8 · Apr 2024

Consolidates research on highway energy-harvesting technologies — piezoelectric, thermoelectric, pyroelectric, electromagnetic, solar — with a comparative assessment of principles, trade-offs and challenges for researchers, engineers and policymakers.

Published

Real-Time Disease Classification in Coffee Plants with Coral Developer Board Mini

IEEE AGRETA 2024 · DOI 10.1109/AGRETA61912.2024.10948998 · Jul 2024

Automated coffee-leaf disease classification on a Coral Developer Board Mini with Edge TPU; MobileNetV2 reached 98.9% training accuracy, with the Coral board outperforming a GPU workstation on inference latency.

In review

Detecting Coffee Leaf Diseases with the Google Coral Developer Board Mini

Computers and Electronics in Agriculture · status: first correction
In review

Wired Highways: Energy Production for Roads

Open Access Journal of Power and Energy (OAJPE) · status: submitted after first correction

Skills & Tools

04 / STACK

Embedded & Firmware

TI MCUBare-metal CLinux firmware STM32ArduinoKeil uVision

Power, PCB & Automation

Circuit & PCB DesignKiCADPower Electronics Cell Testing & ValidationFanuc / Mitsubishi RoboticsWiring Systems

Data, Simulation & Delivery

LTspiceSimulinkAnsys Fluent Python / MLTroubleshootingCross-functional Delivery

Education & Certifications

05 / CREDENTIALS

MEng, Electrical & Electronics Engineering

University of Nottingham · 2019 – 2024

Foundation in Engineering

University of Nottingham · 2017 – 2018

Languages

  • English Professional
  • Hindi Native
  • Urdu Native
  • Malay Intermediate
  • C Language, Basic to Advance Learnvern, ongoing
  • Legacy JavaScript Algorithms & Data Structures freeCodeCamp, 2023
  • Data Analysis with Python freeCodeCamp, 2022
  • Machine Learning with Python freeCodeCamp, 2022
  • Responsive Web Design freeCodeCamp, 2022
  • Data Visualisation using Python IBM, 2021
  • Big Data 101 IBM, 2021
  • Computer Networking — Wired & Wireless Alison, 2021
  • Python for Pentesters Alison, 2021