My work is organized around three pillars. In each one, I detail the tools and methods I use to design, test, and improve my embedded projects.
Good software is simple, readable, documented, and reproducible. These are the principles I apply to build a solid software foundation in my embedded projects.
Prototypes are modular and scalable to enable progressive integration of ideas at lower cost.
Each project follows a hybrid agile development method, encouraging creativity and responsiveness while ensuring the need is addressed.
Good software is simple, readable, documented, and reproducible. These are the principles I apply to build a solid software foundation in my embedded projects.
For embedded firmware development, I use the libraries provided by microcontroller manufacturers. Software design always starts with a benchmarking step (existing open-source code) to validate technical choices, then continues with step-by-step development of product subsystems (drivers, communication protocols, human-machine interfaces), each with dedicated tests. I also use JTAG and GDB debugging phases to validate the system.
Prototypes are modular and scalable to enable progressive integration of ideas at lower cost.
Numerical simulation is essential to validate solutions. I start by testing functions whenever there is uncertainty. Then I move to breadboards or build a PCB prototype. Prototypes are debugged using an oscilloscope, logic analyzer, and multimeter to validate behavior. They always include extra connectors so the design can evolve as the project progresses.
Each project follows a hybrid agile development method, encouraging creativity and responsiveness while ensuring the need is addressed.
The high-level direction is defined first (objectives, performance). No detailed technical choice is fixed at this stage so there is room for creativity and adaptation.
Using the hybrid agile method, I build a minimal prototype to test the most critical assumptions as early as possible and reduce technical risk.
Hardware and software are integrated progressively, with performance measurements used to validate behavior.
I keep a daily engineering logbook where I record results and feedback to keep the project maintainable and make future iterations easier.
