Research About Projects Posts Contact
bakikocabasa@gmail.com
Ph.D. Candidate at Volkswagen AG

Hi, I'm Baki Kocabasa

Researcher and engineer at the intersection of battery technologies, data science, and computational modeling. Building the future of electric mobility at Volkswagen AG & TU Braunschweig.

My Research Resume
Ph.D. Volkswagen AG × TU Braunschweig
M.Sc. Computational Science in Engineering
Since 2022 Battery research at Elenia & Volkswagen
Research

Why batteries age, read from field data.

Electric vehicles produce battery data every day, but it is noisy, incomplete and never recorded with diagnostics in mind. In my Ph.D. at Volkswagen AG and TU Braunschweig, I develop non-destructive methods that turn this field data into an understanding of why cells age. Combining incremental capacity and differential voltage analysis (ICA/DVA) with mechanistic degradation models, I separate degradation modes such as loss of lithium inventory and loss of active material, so ageing can be traced to its root cause and predicted.

  • Field data, not lab cyclingDiagnostics from the measurements vehicles already record.
  • Mechanisms, not a single numberInterpretable degradation modes instead of one State-of-Health value.
  • Built to scaleMethods designed for fleet-sized data and battery quality work.
Schematic incremental capacity curves Two dQ/dV curves over cell voltage. The aged curve's peaks are lower and shifted to higher voltage compared with the fresh curve. Cell voltage dQ/dV Fresh cell Aged cell peaks shift and shrink
Illustration of incremental capacity analysis (ICA). How the peaks move and shrink with age points to the underlying degradation mode.
Baki Kocabasa
About Me

Driven by curiosity.
Powered by data.

I'm a Ph.D. candidate at Volkswagen AG in collaboration with TU Braunschweig, focusing on battery field data analysis and the development of non-destructive, mechanistic models to identify the root causes of battery degradation. Through this work, I aim to predict, understand, and ensure battery quality for the future of the automotive industry. My research lies at the intersection of electrical engineering, method development, and computational modeling.

With a Master's in Computational Sciences in Engineering and a Bachelor's in Electrical Engineering, I bring a multidisciplinary approach to solving complex problems in energy systems and machine learning.

Based in Braunschweig, Germany, I'm passionate about leveraging computational methods to advance sustainable transportation and energy technologies.

Education M.Sc. Computational Sciences (TU Braunschweig)
Focus Battery Data Analysis & Modeling
Location Braunschweig, Germany
Expertise

What I work on

Four areas I work in, each linked to something you can read or run.

Battery Diagnostics & Modeling

Field data analysis, ICA/DVA, mechanistic degradation models and State-of-Health estimation for EV batteries.

ICA/DVA SOH/SOC Degradation Field data
LSTM State-of-Health project

Data Science & Machine Learning

Time-series analysis, feature engineering, clustering and model deployment, from my master thesis on driving-style identification at Volkswagen.

Python scikit-learn Time series AWS
Missing-data guide

Deep Learning & Computer Vision

CNNs, transfer learning, segmentation and object detection for aerial and remote-sensing imagery.

TensorFlow PyTorch YOLOv5 CNNs
YOLOv5 project

Energy Systems & Numerical Methods

Optimization and simulation for energy management: genetic algorithms, numerical methods, MATLAB/Simulink.

MATLAB Simulink Optimization Numerics
EV energy management on GitHub
Projects

Featured Work

Battery and energy work first, followed by deep learning, numerical methods and a side project.

Capacity fade of NASA Li-ion cells over charge-discharge cycles
Battery

Battery State-of-Health Prediction with LSTMs

Capacity-fade analysis of NASA Li-ion cells and stateful LSTM models that predict State of Health across charge–discharge cycles.

Read Case Study
Energy Management System
Optimization

EV Energy Management with Genetic Algorithms

Genetic algorithm-based optimization for electric vehicle energy management. Balancing performance, efficiency, and battery longevity.

View on GitHub
YOLOv5 Object Detection
Deep Learning

YOLOv5 Object Detection on DIOR Dataset

Custom-trained YOLOv5 model for aerial image object detection. Full pipeline from data preparation to evaluation.

Read Case Study
Numerical Methods
Numerical Analysis

Numerical Methods for Value Approximation

Implementation of various numerical methods for engineering value approximation problems. Comparing convergence rates and efficiency.

View on GitHub
Family Tree App
Web App

Interactive Family Tree Application

Bilingual (English/Turkish) interactive family tree visualization. Privacy-focused, offline-capable, built with D3.js.

Explore App
Blog

Selected Posts

Write-ups of my projects, plus course notes on machine learning and deep learning.

Deep Learning

Custom Object Detection with YOLOv5

Training YOLOv5 on the DIOR remote sensing dataset for aerial image object detection.
Battery Course project

Predicting Battery Life with LSTMs

NASA battery dataset, capacity degradation analysis, and stateful LSTM architectures for State of Health prediction.
Machine Learning Course project

The Complete Guide to Handling Missing Data

Detection, visualization with missingno, deletion strategies, imputation methods, KNN imputation, and the EM algorithm.
All 22 posts
Get in Touch

Let's work
together.

Whether you're interested in collaboration, have a question about my research, or just want to connect — I'd love to hear from you.

bakikocabasa@gmail.com LinkedIn GitHub Instagram

Email me directly

The quickest way to reach me. I read every message and usually reply within a few days.

bakikocabasa@gmail.com

Prefer LinkedIn? Message me there.