Business Data Scientist
Salomonsen Juul
Business Data Science profile with a commercial mindset, where academic theory meets business understanding. Data should be structured and evaluated so it creates real bottom-line value — not just technically correct results. I've learned to treat evaluation as a fixed part of the work, so I can be honest about what a solution can actually do and where its limits lie. A strong collaborator who thrives in iterative processes where the whole team's competencies come into play.
"Business and data science aren't two separate skill sets to me — they're one discipline, shaped by real projects with real organisations. I do my best work close to the people who'll use it and the colleagues building it with me."
— Mathias Salomonsen Juul
Master's thesis. Synthetic Danish PSOAP clinical note generation using GPT-4o, evaluated through a downstream patient diagnosis pipeline within a Design Science Research framework.
Developed in collaboration with Esben Münster Graahede.
Transforms academic PDFs into two-host podcast audio — Llama 3.1 for script generation, Suno Bark TTS for synthesis, plus a knowledge graph and quiz generator.
Automated weekly pipeline fetching sustainability news, summarising with Gemini 2.5, and generating strategic business consultations — orchestrated by GitHub Actions.
End-to-end analysis of the Kiva microfinance dataset — data cleaning and EDA, K-means and hierarchical clustering, an SVD recommendation prototype, and regression models (Linear Regression, Random Forest, XGBoost) with SHAP explainability.
RAG chat application for internal documents — OpenAI embeddings in ChromaDB, GPT-4o grounded generation, and a FastAPI backend.
Communication network from the Enron dataset — PageRank, betweenness centrality, community detection, LDA topic modelling, and an interactive PyVis graph.
End-to-end luxury e-commerce portfolio — FastAPI backend, SQLite database, synthetic data pipeline, KMeans/RF/GBM ML models, and a static BI dashboard on GitHub Pages.
Granger causality analysis of Danish renewable energy vs. electricity spot prices — VAR modelling, impulse response functions, and a Dockerised FastAPI results service.
Two-stage Gemini pipeline classifying customer region and emotional type, retrieving behavioural context, and generating a personalised outreach message in a tradesperson's voice.
Multi-agent pipeline answering ambiguous business questions against a synthetic dataset — retrieval, analysis, and a cross-provider critic agent (GPT-5.4-nano + Claude Haiku 4.5) with grounding, confidence, and human-in-the-loop guardrails, backed by 75 passing tests.
Formal collaboration providing access to validated clinical knowledge for the CDSS system — carried out on behalf of Danske Regioner and Lægeforeningen.
Master's Thesis · AAU
Structured interviews with doctors and nurses to gather domain knowledge and employee perspectives — directly shaping the requirements and evaluation criteria of the CDSS clinical NLP system.
Master's Thesis · AAU
Final bachelor's project conducted in collaboration with Shaping New Tomorrow Holding ApS — applying business analytics and organisational analysis to a real company context.
Bachelor Project · AAU
Semester project carried out in partnership with Neogrid Technologies ApS — combining data analysis and business strategy in a live technology company setting.
5th Semester Project · AAU
Facilitated consumer interviews across multiple marketing and value-focused projects — translating consumer insights into analytical frameworks and business-oriented recommendations.
Consumer Research · Bachelor's Programme
Primary language across all projects — from data pipelines and statistical modelling to REST APIs and LLM orchestration. Comfortable with the full ecosystem: pandas, NumPy, FastAPI, scikit-learn, and more.
End-to-end data work: ingestion, cleaning, transformation, and exploratory analysis. Experience with structured tabular data, time series, text corpora, and graph data — always focused on reliable, reproducible outputs.
Applied statistical reasoning underpins the work — hypothesis testing, stationarity analysis, Granger causality, VAR modelling, regression, and probabilistic modelling. Equally comfortable with theory and implementation.
Rapid, end-to-end system building — from notebook proof-of-concept to deployed application. Includes API design, containerisation with Docker, GitHub Actions automation, and lightweight frontends.
Relational database design and querying for analytical and transactional use cases. Experience with SQLite and SQLAlchemy ORM — structuring schemas, seeding data, and integrating databases into backend services.
Hands-on with the modern LLM stack — OpenAI, Gemini, and open-source models via Hugging Face. Covers prompt engineering, RAG pipelines, function calling, vector databases, and evaluation workflows across multiple projects.
Cand. merc. Business Data Science
Aalborg University
Master's programme combining business strategy, statistics, and software engineering. Thesis in clinical NLP: synthetic Danish PSOAP note generation and evaluation within a Design Science Research framework.
HA — Business Administration
Aalborg University
Foundation in business economics, organisational theory, and quantitative methods — the domain context that grounds the data science work.
Higher Commercial Examination (HHX)
Business College
Upper secondary education with a business and economics focus — providing the commercial foundation that underpins subsequent academic and professional development.
Bestyrelsesmedlem
Gade og Juul A/S
Board member at Gade og Juul A/S — contributing strategic oversight and business perspective alongside ongoing academic studies in Business Data Science.
Retail Employee
Meny
Part-time retail work alongside studies — developing practical customer service and operational experience throughout the bachelor's and master's programmes.
Retail Employee
Løvbjerg & Føtex
Full-time retail employment during a gap year — building work ethic and commercial awareness prior to commencing university studies.
I hold a cand. merc. in Business Data Science from Aalborg University, at the intersection of analytical methods, software engineering, and commercial thinking — where academic theory meets business understanding. For me, structuring data is only half the job; evaluating it honestly, so I can speak plainly about what a solution can do and where its limits lie, is just as important.
My project work spans the full stack — statistical modelling, machine learning, LLM pipelines, and production deployments — with real industry collaborations including Neogrid Technologies, Shaping New Tomorrow, and a formal partnership with Lægehåndbogen & Patienthåndbogen for my clinical NLP thesis.
Throughout both my bachelor's and master's, almost every project has been carried out in direct partnership with an organisation rather than in the abstract — working with real stakeholders, real constraints, and real business questions. My cand. merc. sits deliberately at the intersection of business and data science, a combination reinforced by my HA in Business Administration and Economics, which gave me the commercial grounding the technical work builds on.
My entire education at AAU has been group-based and problem-driven, giving me a strong collaborative instinct and the ability to take co-responsibility for a team's progress and ensure knowledge flows freely between everyone involved. My thesis in clinical NLP brings it all together — rigorous methodology, real stakeholder collaboration, and practical system design in a high-stakes domain.