Sebastian Speekenbrink
Software & embedded engineer.
Building production-grade infrastructure to support AI-workloads.
A short note.
Amsterdam-based AI Infrastructure engineer & consultant, working day-to-day with complex on-premise, air-gapped Kubernetes clusters. Comfortable across the programming stack — from low-level C, C++ and Rust to Python. Furthermore proficient in cloud platforms (AWS, GCP), containerisation and orchestration (Docker, Kubernetes), and the tooling around them (Ansible, Helm, Terraform).
My theoretical background is in electrical engineering (Bsc) and embedded systems (MSc), having studied at the TU Delft, with an exchange at the University of Queensland in Brisbane, Australia. My focus has been at integrating machine learning models into embedded systems, with a thesis in End-to-End Embedded Machine Learning for In-Ear PPG Peak Detection.
Alongside working as an AI infrastructure engineer & consultant, I do freelancing embedded software engineering for several electrical-vehicle projects.
What I do.
I design and deploy on-premise infrastructure for an AI workload orchestration platform, mostly on
Kubernetes. Consultancy is furthermore provided for the different AI-applications
running on top of the provided stack.
The work cuts through all different layers across the stack: from virtualising AI-servers and
provisioning multi-node clusters to developing user-facing applications.
This all is often performed in complex, air-gapped environments where the public cloud isn't an option.
Embedded software for several electrical-vehicle projects, ranging from circuit design to embedded firmware deployment on MCU's.
A few other things.
I'm an avid sports player, with a background in international-level judo. Currently, my physical activity is mostly focused on running, judo, BJJ and swimming, although you can find me regularly in the gym as well, with my holidays spent surfing.