In my own words

After 10+ years working as a quantitative Portfolio Manager in the Asset Management and Financial industry, in 2018 I decided to make a change and started to study Computer Science. The focus in my studies were on Data Science and Machine Learning.

In 2020, I joined Commerzbank AG to work as a Data Scientist (working student) in their Big Data & Advanced Analytics department. In 2024, I joined adesso SE and have since been working as a Machine Learning engineer and Python developer.

Although I had previously developed and applied computer programs for securities analysis and fund management, I remain fascinated by the current AI revolution and am eager to explore its expanding opportunities for data analysis and beyond.

Education

M.Sc. in Computer Science

2021 - 2024
Frankfurt University of Applied Sciences, Frankfurt

Full time Master degree studies in Computer Science with a chosen focus on Machine Learning and Data Science. I learned the theoretical background of Machine Learning and its algorithms such as Logistic Regression, SVM, Ensembles (Random Forests, XGBoost, etc) and Artificial Neural Networks (ANNs). I extended my knowledge to other Machine Learning fields such as NLP and Computer Vision. I applied this knowledge in multiple Machine Learning projects, gaining further experience with Python and related Data Science and Machine Learning libraries.

  • Final Grade: 1.3 (top 5%)
  • (Optional) Modules chosen: Machine Learning, Data Mining, Learning from Data, NLP Project, Speech Recognition, Computer Vision, Adaptive Knowledge Systems
  • Multiple coding projects using Python and related libraries: PyTorch, scikit-learn, spaCy, pandas, NumPy, OpenCV,etc.
  • Master-Thesis: “Constructing a Knowledge Graph by extracting information from financial news articles”

B.Sc. in Computer Science

2018 - 2021
Frankfurt University of Applied Sciences, Frankfurt

Full time Bachelor degree studies in Computer Science. I refreshed and enhanced my knowledge in mathematics and learned the theoretical concepts of Computer Science. I gained hands-on programming experience in different programming languages during multiple projects. I particularly enjoyed practical modules such as Algorithms & Data Structures, Databases, and Computer Networks. My Bachelor-Thesis focused on Machine Learning topics and feature contribution techniques such as Shapley values.

  • Final Grade: 1.6
  • Bachelor-Thesis: Title: “An approach to Out-of-Model feature explanations in the application of credit fraud detection”

M.A. in Economics

1994 - 1995
University of San Francisco, CA, USA

Full time Master degree studies in Economics with a focus on quantitative subjects that was partly funded by merit scholarships due to outstanding grades. I also worked as a Teaching Assistant, conducted exercises and designed, corrected and graded undergraduate exams.

  • GPA: 3.875 (top 5%)
  • Focus: Statistics, Econometrics, Macroeconomic Modelling
  • Master-Thesis: Title: “The macroeconomic aspects of German Reunification”

Pre-Diplomas in Economics, Law, Political Science

1991 - 1994
Ruprecht-Karls University, Heidelberg

Preliminary diplomas in three different degree studies. My main interests were Macroeconomics, Economic Policy and Public Law.

Experiences

Machine Learning Engineer - Python developer

Mar 2024 - Today
adesso SE, GenAI-department, Data & Analytics, Frankfurt

Machine Learning engineer and Python developer on multiple AI projects. In this role I:

  • built data extraction and processing pipelines using fine-tuned Azure Document-Intelligence and OpenAI models
  • built data and image extraction and processing pipelines using pure Python and Python libraries
  • built LLM processing pipelines using LangChain and LangGraph
  • extracted structured information from PDFs into a Neo4j Graph to allow customized LLMs to query related information (Alternative RAG)
  • built MCP servers for agentic networks
  • advised clients on different Prompt Engineering techniques
  • reengineered a full stack web app (backend, frontend)
  • wrote CI/CD pipeline code to containerize and deploy models to Azure

Data Scientist (part time)

2020 - Feb 2024
Commerzbank AG, Big Data & Advanced Analytics, Frankfurt

I have worked as a part time Data Scientist (working student) on two main projects:

  • Machine Learning model to detect corporate credit fraud
  • Large Language Models (LLMs) to classify the text of News articles

For that, I:

  • used Python libraries (pandas, re, spaCy, etc) to preprocess and prepare text and other unstructured and structured data.
  • applied traditional algorithms such as Logistic Regression, SVM, Random Forests, XGBoost and optimized the model’s hyper-parameters.
  • downloaded, fine-tuned and applied HuggingFace models to classify the text of News articles.
  • worked with: Cloudera Data Science Workbench (CDSW), Google Cloud Platform (GCP), Docker, Kubernetes, Hadoop ecosystem, Git/Bitbucket, Confluence, etc.

Consultant, CEO

2014 - Spring 2024
Portfolio-Resolution, Frankfurt

I advised institutional investors, particularly on their ABS/CDO portfolios. I analyzed securities, verified price marks and helped clients to meet regulatory requirements. I also advised clients on their portfolio holdings and made asset allocation recommendations.

Portfolio Manager - Entrepreneur

2013 - 2014
LeanVal Asset Management (formerly CCPM AG), Frankfurt

Entrepreneurial attempt and preparation to set up a quantitatively managed credit hedge fund using credit derivatives. The plan eventually failed as the main seed investor unexpectedly withdrew his initial commitment.

Head of Asset Management - Credits

2007 - 2012
Oddo BHF Asset Management (formerly Frankfurt-Trust), Frankfurt

I steered a team of seven Portfolio Managers that managed funds with a combined volume of EUR 1.3 Bio. through the financial crisis of 2008. I preemptively sold particularly risky securities before their market prize plunged. I implemented quantitative investment approaches to analyze securities and manage investment funds more effectively. I implemented infrastructural processes (E-Trading, Back-Office, etc.) to reduce administrative burdens for Portfolio Managers. I promoted the company and our fund management concepts, winning prestigious management mandates from institutional investors.

Head of Asset Management - Credits, Portfolio Manager

2000 - 2007
W&W Asset Management, Stuttgart

I led a team of three Portfolio Managers that managed two dedicated credit funds with a combined volume of EUR 800 Mio. I developed a comprehensive concept to manage these funds and analyze securities quantitatively. I implemented it by writing code in Excel Visual Basic (VBA). The two funds won many awards and prizes for their outstanding performance:

  • W&W Euro Corporate Bond Fund
  • W&W Asset-Backed Securities (ABS) Fund

In 2000/2001, I also received the CFA and DVFA/CEFA designations from the CFA Institute and the DVFA Deutsche Vereinigung für Finanzanalyse und Asset Management e.V.

Github Repositories

My projects are hosted on Github. Click links below for some of my public repos or extracts thereof:

UASFRA-MS-MasterThesis - Python/Jupyter Notebook to my Master-Thesis (2024) in Computer Science. Extract named entities, resolve coreferences and model topics of news article sentences to be stored in a neo4j Knowledge Graph. The Knowledge Graph is enriched with data from Wikidata, DPPedia data. It can be queried by an OpenAI chat bot.
UASFRA-MS-KnowledgeGraph - Python project to read and use ESG data from XBRL-files to construct a neo4j Knowledge-Graph to be enriched with external data (Wikidata, DBPedia). An OpenAI-attached chat bot is used to query the Graph.
UASFRA-MS-ProjektIntellSys - M.Sc. Computer Science project in cooperation with www.right-basedonscience.de - Jupyter notebook/Python scripts to extract ESG related KPIs from PDF documents.
UASFRA-MS-LFD-FINETUNING-DEMO - Demo Jupyter notebook/Python scripts to download a pretrained language model from Hugging Face and fine-tune it according to your own topic domain and needs.
UASFRA-BS-BachelorThesis - Explaining contributions of features that are not part of a Machine Learning model by using Transfer Learning and Shapley values/SHAP.
NLP-CleanText - Part of a larger NLP Machine Learning project. Python scripts to clean and preprocess raw, unprocessed, and messy text, mostly using regular expressions (Python re package, Python 3.11).

Publications

My publications as a Computer Science student can be downloaded as PDF.

Constructing a Knowledge Graph by extracting information from financial news articles
Rainer Gogel
Master-Thesis, Frankfurt University of Applied Sciences, M.Sc. Computer Science, 2024
Msc. Computer Science Project: Knowledge Graph construction based on XBRL data
Rainer Gogel
Frankfurt University of Applied Sciences, M.Sc. Computer Science, Project, 2024
Talk 15: Transformer Applications in NLP: An explanation of BERT
Rainer Gogel
Frankfurt University of Applied Sciences, M.Sc. Computer Science, Learning from Data, 2023
An approach to Out-of-Model feature explanations in the application of credit fraud detection
Rainer Gogel
Bachelor-Thesis, Frankfurt University of Applied Sciences, B.Sc. Computer Science, 2021

Skills & Proficiency

Python

pandas, NumPy, re

scikit-learn

PyTorch

spaCy

LangChain

Docker

matplotlib, seaborn, plotly

OpenAI

sql, mysql, peewee ORM

Neo4j

Azure

HuggingFace

LangGraph

Cloudera Data Science Workbench (CDSW)

hadoop ecosystem

GCP, AWS

html, css

Kubernetes

javascript, typescript

Java, C, C++