Jorgen Bergh

Jorgen Bergh

ML engineer and student at NTNU and UC Berkeley, building systems where machine learning meets real-world constraints.

Trondheim / Oslo / Berkeley

jrg.bergh@gmail.com

Jorgen Bergh

(01)About

I study Industrial Economics and Technology Management at NTNU, specialising in Computer Science with a focus on AI and ML. In 2025–26 I spent a year as a visiting student at UC Berkeley, where I took courses in machine learning, economics and financial engineering.

My work spans industries with high stakes and hard constraints — particle physics trigger systems at CERN, real-time monitoring across Norway's hydropower grid, and sustainable AI for e-commerce. As part of COMPSCI 294 at Berkeley, I co-authored a preprint on carbon-aware recommendation re-ranking that was published on arXiv.

Outside of work I enjoy staying active through weight training, endurance training, and basketball. I competed in both athletics and basketball through upper secondary school.


(02)Experience
Edda AI

AI Engineer Intern

Aug–Sep 2026 · Oslo, Norway
CERN Openlab

ML Intern

  • FPGA implementation of a VAE encoder for the ATLAS trigger using hls4ml and quantization-aware training (HGQ2), meeting strict latency and on-chip resource budgets.
  • Chose this position over an offer from Rystad Energy as an Analysis/AI Intern.
Jun–Aug 2026 · Geneva, Switzerland
Å Energy

AI/ML Intern

  • Built PyTorch models for digital monitoring and anomaly detection across 82 hydropower plants.
  • Applied language models and computer vision to enhance decision support in hydropower operations.
Jun–Aug 2025 · Drammen, Norway
ReLU NTNU

Project Lead → Data Scientist

  • Led market forecasting for the sonar industry by analyzing ~5,000 scraped news articles with LLM-based sentiment.
  • Built anomaly detection model on energy time series from ~200 buildings with five years of hourly readings.
Jan 2024–Aug 2025 · Trondheim, Norway

His Majesty the King's Guard

Second-in-Command

  • Responsible for overseeing up to 30 soldiers during active duty and military training.
Oct 2022–Sep 2023 · Oslo, Norway

(03)Education

NTNU — Norwegian University of Science and Technology

M.Sc. Industrial Economics and Technology Management

Computer Science (AI/ML) + Economics/Finance. GPA: 4.4/5.0

University of California, Berkeley

Visiting Student

GPA: 3.97/4.0

Fall 2025

  • COMPSCI 188 — Introduction to Artificial Intelligence
  • DATA C100 — Principles & Techniques of Data Science
  • INDENG 221 — Introduction to Financial Engineering
  • UGBA 101A — Microeconomic Analysis for Business Decisions
  • BUSADM 460.1 — Marketing Management
  • INDENG 198 — Introduction to Quantitative Finance
  • SOCIOL 198 — Directed Group Study for Undergraduates

Spring 2026

  • COMPSCI 189 — Introduction to Machine Learning
  • COMPSCI 288 — Advanced Natural Language Processing
  • COMPSCI 294 — Designing Algorithmic Media (A+) — course project published as arXiv:2606.04550
  • ECON 141 — Econometrics (Quantitative) (A+, best final exam score)
  • ENGIN 183E — Technology Entrepreneurship
Jorgen and his group pitching Stratum at ENGIN 183E
Me and my group after pitching Stratum — our AI automation platform for ESG reporting — in front of VCs and industry guests as part of ENGIN 183E (Technology Entrepreneurship). Also pictured: our professor Naeem Zafar.

(04)Selected Work
CS 294 — Designing Algorithmic Media
Carbon-Aware Re-ranking for E-commerce Recommendations

Course project for COMPSCI 294 (Designing Algorithmic Media) at UC Berkeley, which became a co-authored arXiv preprint. Explores trading user engagement for sustainability in e-commerce recommendation systems.


Hardware ML

ATLAS Trigger VAE — CERN Openlab

FPGA implementation of a variational autoencoder for real-time particle physics event selection using hls4ml.


Time Series

Hydropower Anomaly Detection — Å Energy

PyTorch models for digital monitoring of 82 hydropower plants using time series and computer vision.


Hackathon
CogniView — Railtracks Multimodal Hackathon

1st place (best use of Railtracks). Multimodal AI application built at the Railtracks Hackathon with Noah.


(05)Competitions
  • 1stDigitalOcean Hackathon
  • 1stRailtracks Multimodal Hackathon
  • Best modelRystad Energy Case Competition
  • 2ndTzafon Hackathon
  • 5th (U23 team)Norwegian AI Championship
Railtracks Multimodal Hackathon
From left: Noah (my teammate), me, and Logan (Railtracks representative) — after Noah and I won best use of Railtracks at the Railtracks Multimodal Hackathon.

(06)Writing

Occasional posts on machine learning, things I'm learning, and whatever else I find worth writing about.

Read the blog →

(07)Interests

Training and health have been a core part of my life for as long as I can remember. I stay active through weight training, endurance work, and basketball — and I track my progress with regular benchmark tests.

Outside of training I'm into poker, philosophy, dog training, and small game hunting. I've also been part of Poker@Berkeley (Berkeley's poker club), NTNUI Basketball, and NTNUI Volleyball.

Jorgen playing basketball

(08)Skills

Technical

Python, PyTorch, TensorFlow, SkLearn, Pandas, C++, Java, SQL, KQL, Azure, HLS, Linux

Languages

Norwegian (native) · English (fluent) · Spanish (conversational)