Skip to content

Dr. Thomas R. Holy

Machine Learning Specialist · Postdoctoral Researcher · Statistical Learning

📧 info[at]thomas-robert-holy.de · 🌐 LinkedIn · 💻 GitHub · 📍 Jena, Germany

Skills

Methods

  • Statistical learning: regularized regression (LASSO, Elastic Net, Ridge), support vector regression, robust losses (Huber, quantile, ε-insensitive), tree ensembles, autoencoders
  • Feature selection: recursive feature elimination, sparsity-inducing penalties
  • Optimization: convex and mixed-integer programming, constrained portfolio optimization (tracking error, cardinality, turnover, UCITS 5/10/40)
  • Quantitative finance & risk: index tracking, weather derivative pricing
  • Forecasting: energy load forecasting
  • Information retrieval & LLMs: hybrid search (dense embeddings + BM25), retrieval-augmented generation, local LLM inference; LoRA fine-tuning of diffusion models

Engineering

  • Languages: Python (primary), R, MATLAB, Bash, SQL
  • ML & data stack: scikit-learn, XGBoost, PyTorch, Optuna, pandas, Polars, NumPy, statsmodels
  • Optimization stack: CVXPY with Gurobi, CPLEX, Clarabel, ECOS, SCS
  • Visualization & apps: Matplotlib, Seaborn, Plotly, Streamlit, Plotly Dash, Flask
  • Backend & data: FastAPI, PostgreSQL, Qdrant, S3/MinIO object storage
  • MLOps & DevOps: Docker / Docker Compose, Singularity, GitLab CI/CD, pytest, ruff, MkDocs documentation, scheduling & monitoring (cron, Gotify)
  • LLM tooling: Ollama, Open WebUI, Flowise, n8n, ComfyUI
  • Infrastructure & HPC: Slurm, Proxmox VE, GPU passthrough, Linux server administration
  • Methodology: CRISP-DM, MLOps

Languages

  • German: native
  • English: full professional proficiency

Papers

Awards & Honors

Selected Open-Source Projects

  • SeFuSe – Privacy-preserving hybrid semantic search over German and EU funding databases (FastAPI, Ollama, Qdrant, PostgreSQL, Docker)
  • dwdown & dwdown dashboard – Weather data pipeline for DWD forecasts with S3 storage, integrity checks and automated recovery

Professional Experience

Machine Learning Specialist

Stadtwerke Jena Gruppe
May 2026 – Present

Developing an end-to-end machine learning workflow for energy load forecasting. Designing a pipeline that covers data ingestion, data pre-processing, model training, evaluation, and serving results for further investigation and process optimization. The system is implemented as a set of modular, containerized services, so that individual components can be deployed and updated independently.

Research Associate – Zentrum für Angewandte Künstliche Intelligenz (ZAKI)

Ernst-Abbe-Hochschule Jena
October 2023 – Present

Contributing to ZAKI (Center for Applied Artificial Intelligence), a platform that brings together stakeholders from academia, business, and civil society to support the practical application of artificial intelligence in Thuringia.

Research Associate / Doctoral Student – Chair of Business Administration, especially Finance

Friedrich Schiller University Jena
April 2021 – September 2023

Dissertation focused on applying machine learning to financial index tracking, developing an interdisciplinary approach to replicate equity indices through optimized sampling techniques. Alongside research, taught in courses on sustainable asset management, derivatives, and introductory economics, including project-based sessions on portfolio selection and stock options, as well as seminars on academic research methods that used exchange-traded funds as a case study. Supervised seminar papers and theses, and supported the editorial management of the academic journal Review of Managerial Science.

Education

PhD in Business Administration (Dr. rer. pol.)

Friedrich Schiller University Jena
April 2021 – November 2023

Dissertation: Financial index tracking with machine learning: An interdisciplinary approach to replicate equity indices using optimized sampling

Master of Science in Accounting, Taxation, and Capital Markets

Friedrich Schiller University Jena
October 2018 – March 2021

Thesis: Smart Indexing with Data-Driven Methods: Theoretical Foundations and Empirical Findings

Bachelor of Science in Business Administration

Friedrich Schiller University Jena
October 2015 – September 2018

Thesis: Management of Weather Risks with Weather Derivatives: Economic Relevance, Construction, and Evaluation