Felix Seitzer

Technical Skills:

Python, Data Science (NumPy, Pandas, PyTorch), Git, C++ (Object-oriented)

Education

M.Sc., Artificial Intelligence, Ostbayerische Technische Hochschule Amberg-Weiden (Oct 2024 - Mar 2026)
B.Eng., Renewable Energy Engineering and Energy Efficiency, Ostbayerische Technische Hochschule Regensburg (Oct 2019 - Mar 2024)
Technisches Fachabitur, Fach-Ober-Schule Regensburg (FOS) (2016 - 2019)

Work Experience

**AI Developer & Data Scientist @ TIKI GmbH (Sep 2024 - Mar 2025)**

At TIKI (Technologisches Institut für angewandte Künstliche Intelligenz GmbH), I worked as an AI Developer & Data Scientist on a project utilizing Deep Learning-based image classification.
My contributions included:

Developing a small and efficient model to significantly improve accuracy.
Optimizing deep learning architectures for real-time image classification.

**Working Student & Intern @ deXcon GmbH (Aug 2022 - Jun 2023)**

deXcon GmbH develops software, components, and control cabinets for decentralized power generation, focusing on P/Q control, remote connections to control centers, and integration with virtual power plants for electricity marketing.

My tasks:

Automation of device configuration via Python
Mechanical assembly of control cabinets
Configuring data loggers and power meters

Projects

Bachelor Thesis

Title: Short-term Photovoltaic Power Forecasting using Deep Learning

In this thesis, I developed a Deep Learning model for short-term PV power forecasting. The model, an end-to-end CNN-MLP hybrid network, takes a sky image, PV historical data, and other data as input to predict future power values for a 15-minute forecast period.

High-level overview of the model:
model architecture

For training and evaluation, a dataset of sky images and PV power values was created using cost-effective software and hardware components such as a Raspberry Pi.

Benchmark Model Implementation

I created a Python implementation of the Smart Persistence Model (SPM) to serve as a benchmark model for the short-term PV power forecasting problem.