Patrick Sheehan / Trondheim, Norway
MSc student in Cybernetics & Robotics at NTNU. Control team at Vortex NTNU. Formerly ML at Cogito NTNU.
01 / About
I'm pursuing a Master's in Cybernetics and Robotics at NTNU in Trondheim, working across machine learning, control systems, and autonomous robotics.
I'm currently on the control team at Vortex NTNU, where I previously served as Project Manager. Before that I spent a semester with Cogito NTNU, the student organization for applied machine learning and AI.
AI tools are part of how I work every day, Perplexity for reading papers, Claude for coding, and all of them for the long tail of small questions that used to cost me hours. I've built a solid foundation through them, across both learning and project work.
Prior to university I served with the Norwegian Armed Forces in HMKG Band & Drill Team, His Majesty the King's Guard.
02 / Experience
Project Manager → Control Team · AUV Software
NTNU's student team building autonomous underwater vehicles for international competition. Served as Project Manager from Feb 2024 – Jun 2025, leading the team across mechanical, electrical, and software disciplines. Currently on the control team, developing the gripper manipulator software stack:
GripperState, GripperWaypoint,
GripperReferenceFilter, and the
GripperReferenceFilterWaypoint action
(vortex-msgs).
vortex-gripper).
vortex-auv.
Machine Learning · Member
NTNU's student-driven organization for applied machine learning and AI. Collaborated on hands-on projects across computer vision, NLP, and reinforcement learning, building practical ML systems alongside a team of driven students.
MSc · Cybernetics & Robotics
Coursework across real-time systems, estimation & classification, modelling & simulation, and embedded systems. Norwegian University of Science and Technology, Trondheim.
HMKG · Band & Drill Team
Completed service with His Majesty the King's Guard (Hans Majestet Kongens Garde) prior to beginning studies at NTNU.
03 / Projects
Manipulator software for Vortex NTNU's AUV, the controller and guidance layers, plus the hardware and simulation interfaces that connect them. End-to-end goal sends now flow through the full pipeline and execute in simulation.
Vortex NTNU
Designed the ROS 2 message and action contracts between guidance, control, and hardware for the gripper, the interface boundary that keeps each layer independent and individually testable.
Vortex NTNU
Fault-tolerant peer-to-peer elevator network. Nodes share state over the network, survive crashes and partitions, and guarantee no order is lost.
NTNU · Real-Time Systems
Nearest-neighbour and k-means classifiers written from scratch, with confusion matrices and error analysis across the full training and test sets.
NTNU · Estimation & Classification
Deriving mathematical models of mechanical, electrical, and hybrid dynamic systems from physical first principles, then solving them numerically and using simulation as an engineering tool for analysis and design, the foundation that control and robotics work builds on.
NTNU · Modelling & Simulation
04 / Self-Hosted
I self-host a local large language model on my own hardware, reachable from anywhere through a secured Cloudflare tunnel. The whole system runs as a set of containerized services coordinated by Docker Compose, with GPU passthrough so inference stays fast on consumer-grade hardware.
Beyond the model itself, this project has given me real hands-on experience with Linux, container networking, reverse tunnels, and the day-to-day work of keeping self-hosted infrastructure running.
05 / Contact