Andreas Bollig, Developer in Aachen, North Rhine-Westphalia, Germany
Andreas is available for hire
Hire Andreas

Andreas Bollig

Verified Expert  in Engineering

Artificial Intelligence (AI) Developer

Location
Aachen, North Rhine-Westphalia, Germany
Toptal Member Since
November 22, 2019

With a Ph.D. in electrical engineering and extensive experience in building machine learning applications, Andreas spans the entire AI value chain, from use case identification and feasibility analysis to implementation of custom-made statistical models and applications. Throughout projects, he stays focused on solving the business problem at hand and creating value from data.

Portfolio

Henkel
Spark, Docker, XGBoost, Scikit-learn, Pandas, Python, Machine Learning, Jupyter
AXA
Scikit-learn, Pandas, Apache Kafka, Spark, Python, Machine Learning, Jupyter
RWTH Aachen University
LaTeX, Python, MATLAB

Experience

Availability

Part-time

Preferred Environment

Databricks, Spark, Scikit-learn, Docker, Python

The most amazing...

...machine learning solution I've built is a fully automated cloud-based end-to-end cost-type proposer for GL bookings.

Work Experience

Senior Data Scientist

2017 - 2019
Henkel
  • Built multiple machine learning applications (classification and time series forecasting) including tech stack selection, architecture, model training, and deployment.
  • Mentored junior data scientists in multiple projects incl. own contributions: marketing analytics based on IoT data, production optimization, accounts receivable prediction, epigenetics research, hair color prediction from sensor measurements.
  • Consulted and performed QA in multiple machine learning projects with external implementation partners-contexts: source-to-pay, intercompany, EDI.
Technologies: Spark, Docker, XGBoost, Scikit-learn, Pandas, Python, Machine Learning, Jupyter

Data Scientist

2016 - 2017
AXA
  • Helped build the inhouse data lab and established data science approaches at the company.
  • Built statistical models for customer churn prediction.
  • Implemented Spark data preparation and pseudonymization solutions on Hadoop.
Technologies: Scikit-learn, Pandas, Apache Kafka, Spark, Python, Machine Learning, Jupyter

Scientific Staff

2011 - 2016
RWTH Aachen University
  • Participated in three research projects and contributed to eight applications for research grants, publishing research papers at top international conferences and journals.
  • Analyzed large datasets and performed distributed Monte Carlo simulations.
  • Supervised students writing theses and working as student research assistants.
Technologies: LaTeX, Python, MATLAB

Data Generalist at Financial Services Startup

Covered the whole range of data-related tasks at a financial services startup, from ingesting data from different sources into a data warehouse to making in-depth data analyses to support business decisions.

TASKS:
• Refactored existing data ingestion pipelines to allow for code reuse + added pipelines for new data sources + introduced type checking, automated testing, etc.
• Created dashboards for business controlling.
• Created ad-hoc data analyses and discussed their interpretation with stakeholders.

Online Shop Product Tests & Statistical Data Analysis

Statistical analysis of online shop click behavior for test products with the goal of making decisions for product portfolio selection and statistical analysis of buying behavior of new customers and customer quality by first product purchased.

Web Application Development

I handled the full-stack implementation of features for an existing web application (front end, back end, and database). This included refactoring the existing app and introducing automated tests to the project, as well as adding database management with SQLAlchemy and Alembic.

Geocoding and Geographical Clustering for Claims Hotspot Identification

The task was to develop a solution for identifying geographical hotspots of insurance claims. The claims data was cleaned, geocoded, and geographically clustered. The clustering required the development of a custom algorithm.

Forecasting for Sales and Operations Planning

The task was to build a sales and operations forecasting system from scratch. Sales forecasts from sales agents needed to be acquired, and gaps needed to be filled with statistical forecasts. The required production capacity in terms of machine hours needed to be derived from the sales forecast.

Inventory Forecast

To provide the controlling department ample time for preparing devaluation numbers for the month-end closing in the context of slow-moving inventory, I built a machine learning solution that forecasts inventory quantities for a big number of materials and makes the data usable via a web-based dashboard.

Cost-type Proposer

When bookings in a company's ERP system carry the wrong cost types, the controlling department gets a distorted view of the company's expenses. To this end, I developed a machine learning solution that proposes a cost-type (GL account) for each booking and provides this data to the controlling department via the company's BI system.
2011 - 2016

Doctor of Philosophy (Ph.D.) in Wireless Communications

RWTH Aachen University - Aachen, Germany

2005 - 2011

Master of Science Degree in Computer Engineering

RWTH Aachen University - Aachen, Germany

JANUARY 2015 - PRESENT

Cambridge C2 Proficiency

Cambridge Assessment English - University of Cambridge

Libraries/APIs

Pandas, Scikit-learn, XGBoost, SQLAlchemy, Playwright, PySpark

Tools

Jupyter, LaTeX, MATLAB, Plotly, Pytest, Apache Beam, Cloud Dataflow, BigQuery, GitLab CI/CD, Amazon Athena, Graphviz

Languages

Python, C, CSS, HTML, JavaScript

Paradigms

Data Science

Storage

Exasol, Azure SQL, SQLite

Frameworks

Spark, Hadoop, Flask, Alembic

Platforms

Linux, Docker, Databricks, Apache Kafka, Xen, TinyOS, Azure

Other

Machine Learning, Artificial Intelligence (AI), Natural Language Processing (NLP), Generative Pre-trained Transformers (GPT), Metabase, Poetry, Web Applications, PyProj, geopy, HyperOpt, MLflow

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring