Hi, I’m Georg Walther and here I blog about all things tech and data that I find interesting.
For lack of creativity, here is a rundown of the tools I use and projects I do. Maybe someday I’ll come up with a more interesting bio.
See more about me on my LinkedIn page.
Data science / machine learning experience
- Rare event (e.g. click, conversion, purchase, email open and reaction, generic user action) prediction from user-level time series data using classical machine learning and deep learning
- User clustering based on web tracking time series data
- Data-driven SEM, display, real-time bidding, and direct marketing optimization
- Time-series anomaly detection and reporting for streamed and batched sensor data
- Agent-based asset trading (reinforcement learning)
- Basket-based recommender systems
- Predictive analytics / forecasting of asset demand based on marketplace user behaviour (big data machine learning)
- Use of high-dimensional acceleration time series data to solve a classification problem
- Causal inference using propensity scores and clustering to estimate treatment effects in high-dimensional acceleration time series data
- Implementation of two-tiered deployment CI/CD process for a Python/Django/wagtail app
- Set up scalable Azure-based infrastructure: Terraform templates for instantiation of managed Kubernetes cluster, database backend and virtual network security; Implementation of Kubernetes deployment and service objects as well as ingress resource and controller
- Ideation and implementation of a modern ReactJS web app for the financial sector
Machine Learning & Data Science
Tensorflow, Keras, PyTorch, Numpy, OpenAI Gym, Pandas, Scikit-Learn, Scipy, Spark MLlib
Presto, Apache Spark, PySpark, Spark SQL, Spark MLlib
Apache Airflow, Ansible, Terraform, Azure Cloud, Google Cloud Platform, Kubernetes, NGINX kubernetes ingress controller, Docker, Docker-Compose, Heroku IaaS, GitLab CI
Apache2, Falcon, Gunicorn, Nginx, PyPy, Pytest, Python, Requests, Django, Wagtail, Flask, Keycloak
AWS, Google Cloud Platform, Azure, Azure Managed Kubernetes Service (AKS), AWS Elastic Container Service (ECS), AWS Elastic Kubernetes Service (EKS)
Frameworks / concepts / methods
Anomaly Detection, Continuous Delivery, Continuous Integration, DevOps, Model Selection (AIC, BIC), Reinforcement Learning, Supervised Learning, Unsupervised Learning, Frequency domain feature engineering (fast Fourier transformation / FFT, discrete Fourier transformation / DFT, short-term Fourier transformation / STFT), Causal Inference