Majd Jamal
I build machine learning systems that help teams forecast outcomes, identify unusual patterns and analyse documents with LLMs. I’m available for B2B consulting assignments in machine learning and data science.
Svenska kraftnät — Sweden's transmission system operator (TSO)
Aug 2026 – present
I develop regression models to forecast imbalances in Sweden's electricity system, drawing on published research.
Riksrevisionen — Swedish National Audit Office
Oct 2022 – Aug 2026
I developed tools to help auditors identify unusual transactions in general ledgers, retrieve relevant documents using RAG and analyse text with LLMs.
Pensel Konst Sverige AB
Dec 2023 – Jun 2025 · founded and built alongside full-time employment
As co-founder and CTO, I built the technology behind an e-commerce service that turned customer descriptions into AI-generated artwork, delivered as framed prints.
KTH Royal Institute of Technology
Sep 2021 – Oct 2022; 2026 – present
I teach artificial neural networks and probabilistic models to master's and doctoral students, and assess their coursework.
KTH × Saab Aeronautics
Jan – Jun 2022 · Järfälla, Sweden
I evaluated CyCADA for adapting models from simulated to real aerial imagery. The evaluation did not show an improvement in object detection, and I reported that finding in the thesis.
I explain how a neural network learns, using a board and marker to make the mathematics easier to follow.
My projects have taught me that creating value means making technology easy to adopt and access.
2016 – 2022
M.Sc. in Computer Science, specialising in Machine Learning — 2022
PyTorch, TensorFlow, scikit-learn; transfer learning.
Hybrid retrieval with BM25, FAISS and Reciprocal Rank Fusion; embeddings; local LLMs served through Ollama.
Docker Compose; continuous integration (CI) and testing; APIs with FastAPI.
Python, SQL and PostgreSQL.
Fluent in Swedish, English and Arabic; basic Spanish.
