About Projects Experience Contact

Open to opportunities

Mathias Kræmer Eberhardt Sørensen

Engineering Student in AI & Data  ·  Nærum, Denmark

View Work Get in Touch
Mathias Kræmer Eberhardt Sørensen
scroll
About Me

A few words
about myself

I'm an AI & Data engineering student at DTU with a strong passion for building things — from machine learning pipelines to embedded hardware systems.

My background spans computer vision, IoT, and software engineering, shaped by both academic projects and hands-on personal builds. I enjoy working at the intersection of software and the real world.

Currently finishing my BSc at DTU Lyngby, I'm looking for opportunities where I can contribute meaningfully and keep growing. Outside of tech, my hobbies include strength training, running, swimming, and climbing to keep me sharp.

Selected Work

Things I've
built & shipped

01

Stabilizing Gradient Inversion in Federated Learning

Bachelor thesis at DTU Compute investigating whether failed gradient inversion attacks mean the information is absent from shared gradients, or just hard to extract. By selectively masking gradient components before inversion, VGG-13 reconstructions improved from 15.81 to 23.14 dB PSNR on CIFAR-100 and from 14.14 to 22.50 dB on LFW faces, showing that baseline attacks can underestimate privacy risks in federated learning.

PyTorch Federated Learning Privacy HPC
View Project →
02

U-Net NK Cell Segmentation

DTU course project on segmenting Natural Killer cells in fluorescence microscopy images. Implemented an OpenCV baseline (62.49% IoU) and two U-Net models in PyTorch — achieving 80.35% IoU with a supervised ResNet50 encoder and 78.85% with a self-supervised DINO-pretrained ResNet18, using nested cross-validation on 41 annotated images.

PyTorch OpenCV Transfer Learning HPC
View Project →
03

RFID Smart Locker System

Secure smart-locker combining motion detection, RFID access tokens, and PIN authentication with alarm functionality. Three ESP8266 microcontrollers communicate via MQTT to a Node-RED server handling UID validation, PIN verification, event logging, and an admin dashboard.

ESP8266 MQTT Node-RED Embedded C
View Project →
04

AIBot

Python Discord bot integrating multiple ML models: Stable Diffusion for image generation, CLIP for image classification, and Whisper for audio transcription. Managed GPU/CPU resource allocation and built a command system for generating images, GIF sequences, and transcriptions.

Python Stable Diffusion CLIP Whisper
View Project →
05

SOPHTX

HTX final exam project in Python: a linear algebra library with matrix and vector classes featuring operator overloading, matrix–vector multiplication, constant scaling, and type checking. Documented with Doxygen, benchmarked with timing measurements, and results exported to Excel.

Python OOP Linear Algebra Doxygen
View Project →
Education & Experience

Where I've
been & learned

2023 — 2026

BSc Eng. in AI & Data

Technical University of Denmark (DTU), Lyngby

Current programme covering machine learning, data science, software engineering, and embedded systems. Key projects include computer vision research using U-Net architectures and IoT systems with microcontrollers.

July — December 2026

Exchange Semester

The University of Western Australia (UWA), Perth
  • CITS1402 Relational Database Management Systems
  • ENSC3011 Fundamentals of Mining Engineering
  • SSEH1104 Active Leadership 1: Developing Leadership Skills
2019 — 2022

Higher Technical Examination (HTX)

Zealand Business College, Slagelse — Robotics Technology

Specialised in Robotics Technology with a GPA of 10.3. Developed foundational skills in electronics, programming, and technical problem-solving, culminating in the SOPHTX linear algebra library project.

Contact

Let's make
something together

Whether you have a role in mind, a project to discuss, or just want to say hello — my inbox is open.

mathiaskes@eksworld.dk