Engineering

I received my bachelor's degree in Computer Science from the Norwegian University of Science and Technology (NTNU) in 2026. My main endeavour currently is pursuing a master's in AI at Vrije Universiteit Amsterdam. As a computer scientist, my main interests lie in the intersection between AI, language, consciousness and the human mind; fields I hope to pursue as part of my master's thesis and potentially PhD in the future.

Researcher, Programmer, Designer, Creator

Skills

  • Python
  • JavaScript
  • HTML
  • CSS
  • Vue
  • React
  • PostgreSQL
  • SPARQL
  • SQLite
  • C++
  • Clojure
  • RAG
  • AI agents
  • RxJS
  • Java
  • TypeScript

Ardoq

Team AI

I'm currently working part-time at Norwegian company Ardoq on their AI team, where I work on the utilisation of ontologies to identify data hygiene problems and improve the performance of AI agents in the application.

Internship

I worked as a intern at Ardoq in the summer of 2026, where my team developed an experimental entity extraction solution that allows users to extract valid, previously missing data into the application from unstructured text. I had the main responsibility for document preprocessing, backend API, and fullstack integration between the UI and the backend.

Bachelor’s thesis

Along with my working partner Magnus Eik, I wrote my bachelor's thesis on the applications of Retrieval Augmented Generation (RAG) in STEM education. As part of this project, we performed extensive research on various RAG solutions, implemented a knowledge graph-inspired ingestion and retrieval solution and performed various tests with both volunteers and LLM evaluators. The thesis was graded A.

Learning with artificial intelligence — thesis

Open PDF in a new tab

Other projects

Supergloss

The spring and summer of 2026, I worked on a language learning application that allows users to import books and other resources in a variety of languages, read them and keep track of which words they know.

    Supergloss

    Features

    • Dashboard: Supergloss features a dashboard tab where users can view a quick overview of the words they've learned and the works they've read so far.
    • Vocab: the vocab tab displays a complete and searchable list of the words that are registered for the currently selected language. the user can view and edit the status of words, and remove them from their list.
    • Reader: the Reader tab allows the user to upload and read PDF and ePUB documents, change various settings such as automatically marking unclicked words as known on page turn, and more. The user can also create custom translations and view disambiguations for words with several meanings. The user can also mark words for practice.
    • Practice: the user can practice words they have encountered in the form of flashcards. The user can choose to practice automatically recommended words, or specifically the ones they have marked for practice in the Reader.
    • Settings: the user can change the display language, change their native (translation) language, app theme, or delete their account.
    Supergloss vocabulary dashboardOpen app in a new tab

    Minecraft clone

    For fun, I created a Minecraft clone in the Spring of 2026: CraftMine. It features procedurally generated terrain and villages, mineable materials, and a simple crafting system. I also composed a background music track as part of the project.

      Cogito NTNU

      While I studied at NTNU, I was a member of Cogito NTNU: a student organisation dedicated to learning and experimentation with AI. Among other things, I developed an automatic pipeline that generates quiz videos with questions, answers, music and background footage using AI.

        Cogito NTNU

        Computer Vision

        In my first semester as a member, me and my project group worked on developing monofocal depth detection algorithms on behalf of a company partnered with Cogito.

        HypeAI

        In my second semester, I worked on a team experimenting with how AI can be used to generate social media content. Although we didn't reach our initial goal of 50 000 followers, I personally developed a video generation pipeline making use of AI text and music generation to create quizzes. It receives a text input containing information on a given topic and automatically generates quiz questions, answers and background music, obtains background footage from Pixabay and assembles the video. A sample can be viewed on the right.

        TV2

        In my third and final semester at Cogito, me and my group worked on behalf of the Norwegian news broadcaster TV2 on identifying anomalies in user data using AI techniques and dedicated fine-tuned models. I also participated in a Hackathon this semester developing computer vision designed to identify groceries based on images, and our team got second place.