Senior Data Scientist & Software Developer
NKD
(Consumer goods and retail, >10.000 employees)
Keywords
Skills
- Multiple years of experience developping Data Science project ranging from analytics to production ready machine learning systems.
- Expertice with multiple data science libraries including pandas, numpy, scikit-learn, tensorflow, pymc3.
- Development of production ready code in Python using TDD (Test driven development), DDD (Domain Driven Design), SOLID Design Principles and good use of apropriate design patterns
- Development of code directly in containerized environments (Docker) bridging the gap from development to operations - DevOps
- Advocate of Agile methodologies
Tech Stack:
Python, C++, JavaScript, R, SQL, Linux, Git, Docker, Nginx, Jenkins, Ainsible, Power BI, AWS / Azure
Frameworks & Libraries:
TensorFlow, Keras, Django, Flask, PyMC3, Pandas, Scikit-learn, ExpressJS, React, React Native, Redux
How I work and what I can do:
TDD (Test driven development), DDD (Domain Driven Design), Agile / Scrum, Clean Code, Clean Architecture (SOLID Design Principles), Microservices, Design Patterns, Data Structures, Optimization, Dynamic Programming,
Project history
04/2020
-
04/2021
- Improved the expected gross profit by a six digit figure (€) for each coming season by developing a system based on machine learning models and optimization algorithms to predict sales and adjust the quantity of items bought from suppliers accordingly during the planning phase of the season.
- Identified growth and expansion opportunities by analyzing more than 500 million data points, defining performance metrics and creating Power BI dashboards to share results to C-level management and advisory board of the company.
- Assisted the development of a A/B testing framework to optimize the use of promotion campaings through out Germany.
08/2019
-
04/2020
Data Scientist and Internet of Things (IoT) Developer
IBM
As part of the Industry Lab @Watson IoT Center I took the lead on various proof of concept projects were we tested with state of the art technologies to build solutions for our clients.
- Developed a prototype to solve trajectory optimization problems for automated vehicles with reinforcement learning techniques and using computer vision algorithms for obstacle detection.
- Coordinated and took an active role in the development of a React web application to showcase the solution during the International Motor Show Germany (IAA) and present it to internal management including IBM’s CEO as well as to the general public.
- Developed an object detection algorithm running on the edge (Jetson Nano) for a major energy company in Germany.
- Developed a showcase on how to use Reinforcement Learning to solve the Multiarm bandit problem using Thompson Sampling and displaying the solution through a React Dashboard
09/2018
-
07/2019
Data Scientist Consultant
IBM
- Implemented a classification model to automatize a highly complex manual procedure formerly done by 5 PhD scientist at a major Automotive OEM. This solution was delivered as a dockerized REST API that was deployed to the client cloud instance via a web application.
- Achieved a 20% improvement of hardware utilization in a high performance computing cluster for a major automotive OEM by identifying load balancing opportunities and better job scheduling through data analysis and visualization techniques.
09/2017
-
08/2018
Data Scientist - MBA Technical Consultant
BMW AG
(Automotive and vehicle construction)
* Developed a stochastic risk model (Monte Carlo) and used mathematical optimization techniques to steer the sourcing for new vehicles parts within
the supplier network increasing the NPV calculations of new projects while reducing the risk of having export custom fees to the NAFTA region.
the supplier network increasing the NPV calculations of new projects while reducing the risk of having export custom fees to the NAFTA region.
10/2016
-
08/2017
Intern and Master Thesis
BMW Motorsport, Department: Performance and Simulation
Developed a framework using neural networks and a genetic algorithm to predict the behavior of the race car around a given track and optimize the
suspension parameters to achieve the fastest possible lap time.
suspension parameters to achieve the fastest possible lap time.
Local Availability
Open to travel worldwide
