Daniel Lutziger

Major: Data Science · Minor: Information Systems

August 2022 – December 2025 (completed while working part-time)

• Graduated magna cum laude (5.4/6.0).
• Thesis (5.75): “Towards Scalable Foundation Models in Sleep EEG and Polysomnography” (Idiap Research Institute, supervised by Prof. Dr. Manuel Günther, Dr. André Anjos, Dr. Olivier Pallanca). Built on a multimodal set-then-sequence Transformer encoder pretrained with contrastive learning (LOOC + SimCLR) on ~13,000 subjects across ten heterogeneous PSG cohorts. Trained lightweight downstream heads for sleep staging and apnea detection on held-out SHHS data. Systematically evaluated data scaling, loss functions for class imbalance (focal loss, class-weighted CE), modality ablations, and contrastive objectives.
• Master's Project (5.5): “Spotify Timelines” – built a personalized, context-aware music visualization and recommender system. Evolved into the ACM RecSys 2025 publication (see Publications & Awards).
• Selected coursework: Deep Learning (5.0), Advanced Topics in AI (4.5), Reinforcement Learning (4.5), Foundations of Data Science (5.0), Network Science (5.25), Interactive-Visual Data Analysis (5.75, Best Project Award – Impact on Research), Advanced Software Engineering (6.0), Blockchain & Crypto Economics (6.0), Statistics (5.75).

August 2018 – July 2022 (completed while working part-time)

• Thesis (5.5): “Towards Attitude Analysis using NLP and Graph Analytics: A Feasibility Study” (supervised by Prof. Dr. Alexandre de Spindler). Built a full NLP pipeline to classify political bias in ~25,000 scraped news articles (Fox News, USA Today, etc.). Implemented Named Entity Recognition, sentiment analysis (finance-tuned model), LDA topic modeling, and fine-tuned GPT-3 for political orientation classification. Stored extracted entities and relationships in Neo4j; validated with Similarity, PageRank, and Node2Vec graph algorithms.
• Elective focus on Data Science and Analysis.

June 2025 – December 2025

• Trained transformer-based foundation models on large-scale polysomnography datasets (EEG, EMG, EOG) in PyTorch; designed preprocessing pipelines for multi-channel biosignal harmonization across heterogeneous clinical cohorts.
• Implemented and compared self-supervised pretraining strategies (LOOC contrastive loss, SimCLR, combined objectives); evaluated transfer learning vs. training from scratch on downstream sleep staging and apnea detection.
• Conducted systematic ablations on pretraining data scale, modality subsets, and loss functions for class imbalance; reported results with balanced accuracy, Cohen's κ, and AUROC.

Tech stack: Python, PyTorch, Hugging Face Transformers, NumPy, pandas, Matplotlib, Linux (HPC)

October 2022 – October 2024

• Built the “Groundhog Alpha Telemetry Application” in React (web) and React Native (mobile) for real-time sensor data visualization and live monitoring of a tunnel boring prototype.

Tech stack: React, React Native, JavaScript/TypeScript

August 2013 – Present (apprenticeship full-time 2013–2017, part-time 2018–2025, full-time from January 2026)

Product Owner & Test DevOps Engineer (2018–present)

• Internal Tooling: Led a complete rewrite (V2) in Spring and React of an outdated day-end processing management tool: streamlined the UX, enabled operators to handle multiple tasks simultaneously, added file attachments and automated status email notifications. Later integrated V2 into a broader operations dashboard, consolidating SSO, unifying the database, simplifying user management, and reducing infrastructure costs.
• LLM Proof of Concept:Built a Retrieval-Augmented Generation prototype for internal knowledge management using a local LLM. Demonstrated feasibility within strict infrastructure constraints and established the technical groundwork for the department's AI ambitions.
• Workflow Automation PoC: Designed and implemented a prototype workflow engine to assess viability as a scheduler replacement. Delivered a production scaling assessment that was later independently validated by an external team.
• CI/CD & Deployment: Created the initial automated deployment playbook that secured and standardized deployment sequences across environments; adopted and extended by the wider team as the standard deployment mechanism.
• Automation & Scripting: Built Python scripts and cron jobs on Linux to automate repetitive operational tasks (file processing, batch monitoring), significantly reducing manual effort in day-to-day operations.

Apprentice Software Engineer (2013–2017, full-time)

• Built an iOS application in Swift for UHNW (Ultra-High Net Worth) clients enabling advisor selection through a mobile interface; rewrote a client advisor web application during an Angular framework migration; built a customer-facing live chat for ubs.com (jQuery).
• Completed foundational training across Java, SQL, and web technologies; mentored colleagues on new tooling.

Tech stack: Java, Spring, React, Vite, Angular, Swift, jQuery, Python, Ansible, Linux