a Randstad company

Statistik, Data Science, Biopharma

Profil
Top-Skills
Statistiker Data Scientist Data Engineer Datenqualitätsprüfung CMC Statistik Good Manufacturing Practice Python R Timeseries
Verfügbar ab
01.01.2023
Aktuell verfügbar - Der Experte steht für neue Projektangebote zur Verfügung.
Verfügbar zu
20%
davon vor Ort
20%
Einsatzorte

Städte
Biberach (+200km) Stuttgart (+200km)
PLZ-Gebiete
Länder
Ganz Deutschland, Österreich, Schweiz
Remote-Arbeit
möglich
Art des Profiles
Freiberufler / Selbstständiger
Der Experte ist als Einzelperson freiberuflich oder selbstständig tätig.

1 Jahr 11 Monate

2021-01

heute

Demand Forecasting

Forecaster Python AWS
Rolle
Forecaster
Projektinhalte

- Globale Bedarfsvorhersage für Consumer und Business Produkte

- Globale Vorhersage für Ersatzteile

Kenntnisse
Python AWS
Kunde
Kärcher
Einsatzort
Remote
1 Jahr 3 Monate

2020-10

2021-12

Principal CMC Statistician

Principal CMC Statistician CMC Statistik R Python ...
Rolle
Principal CMC Statistician
Produkte
R
Kenntnisse
CMC Statistik R Python Good Manufacturing Practice
Kunde
Boehringer Ingeheim GmbH & Co KG
Einsatzort
Biberach an der Riß
11 Monate

2019-05

2020-03

Revenue & Next Order Forecasting

Data Scientist & Data Engineer R Python Microsoft Azure
Rolle
Data Scientist & Data Engineer
Kenntnisse
R Python Microsoft Azure
Kunde
ZOI GmbH
Einsatzort
Stuttgart
6 Monate

2019-07

2019-12

Process Analysis for Aircraft Manufacturer

Data Scientist
Rolle
Data Scientist
Kunde
ZOI GmbH
Einsatzort
Augsburg
9 Monate

2019-04

2019-12

Sentiment Analysis Data Pipeline in MS Azure

Data Engineer Python Microsoft Azure
Rolle
Data Engineer
Kenntnisse
Python Microsoft Azure
Kunde
EON
Einsatzort
Essen, Ruhr
5 Monate

2019-06

2019-10

NextGen Seq Analysis

Data Scientist
Rolle
Data Scientist
Kunde
Booehringer-Ingelheim
3 Monate

2019-01

2019-03

Anomaly Detection for PVs

Data Scientist R
Rolle
Data Scientist
Projektinhalte


Kenntnisse
R
Kunde
EON
Einsatzort
Essen, Ruhr
5 Monate

2018-05

2018-09

Spare Part Demand Prediction

Projektinhalte

Evaluating classical time series forecasting methods, e.g., Croston methods and other, in spare part demand forecasting on customer data. Further, develop and compare more advanced methods versus the baseline methods for this type of data (intermittent time series)  as, e.g., Hierarchical Bayesian Models and Neural Networks.

Kunde
Trumpf GmbH + Co. KG
Einsatzort
Ditzingen
1 Jahr

2017-10

2018-09

Smart Meter Forecasting & Anomaly Detection

Projektinhalte

Time series forecasting and anomaly detection of Smart Meter data (B2B). High granular electricity consumption data of more than 10,000 devices, collected over several years were analyzed in view of data quality, improving forecast accuracy, and anomaly detection accuracy. Beside classical forecasting methods, machine learning models and Deep Learning were applied to this type of data. Finally, a productive forecasting system was built in Azure

Kunde
E.ON Business Services GmbH
Einsatzort
Essen
5 Monate

2018-02

2018-06

Anomaly Detection / Data Quality for Financial Data

R Anomaly Detection
Projektinhalte

Extend previously developed methods by including techniques for improving outlier detection accuracy (F1 score and ROC and PR AUC). This includes different ensembling strategies as also pre-selecting features by multivariate correlation measures. As financial datasets are in general high dimensional, outlier detection methods should be prone against irrelevant dimensions and relevant dimensions should be delivered to the business analyst (modification of the Lime).

Kenntnisse
R Anomaly Detection
Kunde
Daimler Financial Services AG
Einsatzort
Stuttgart
8 Monate

2016-05

2016-12

Anomaly Detection / Data Quality for Financial Data

Projektinhalte

The customer must manage hundreds of thousand contracts from different countries that he wants to use for prediction. To get a valid prediction, clean data is absolutely necessary and so a data quality process was established. My task was to examine the actual research on outlier detection algorithms (unsupervised case) for mixed-attribute data, choose appropriate algorithms, implement them, and integrate it into a pipeline. Standard algorithms (as e.g. k-NN and LOF) were adapted to the mixed-attribute scenario. Furthermore, new algorithms (yet not publicly available) are brought to executable code. All algorithms were programmed in C++ and R and where necessary parallelized with Intel TBB.

Kunde
Daimler Financial Services AG
Einsatzort
Stuttgart


4 Jahre 5 Monate

2007-05

2011-09

Promotion

Dr., Universität Stuttgart
Abschluss
Dr.
Institution, Ort
Universität Stuttgart
Schwerpunkt

Maschinelles Lernen

4 Jahre 7 Monate

2002-10

2007-04

Mathematik

Diplom, Universität Stuttgart
Abschluss
Diplom
Institution, Ort
Universität Stuttgart

Certifications and Training

  • SCRUM Master Training

Education

09/2011:

Ph.D. at Institute for Stochastics and Application (Stuttgart)

Machine learning related topic

04/2007:

Diploma in Mathematics with Minor in Mechanics (Stuttgart)

Industry related Diploma thesis (Robotics)

Deutsch Muttersprache
Englisch Verhandlungssicher

Top Skills
Statistiker Data Scientist Data Engineer Datenqualitätsprüfung CMC Statistik Good Manufacturing Practice Python R Timeseries
Schwerpunkte
Machine Learning
Experte
Time Series Forecasting
Experte
Statistik
Experte
Produkte / Standards / Erfahrungen / Methoden
AWS
Fortgeschritten
Microsoft Azure
Basics
Programmiersprachen
Python R

  • Automotive
  • Energy
  • Logistics, and
  • Healthcare.
Ihr Kontakt zu Gulp

Fragen? Rufen Sie uns an +49 89 500316-300 oder schreiben Sie uns:

Jetzt bei GULP Direkt registrieren und Freelancer kontaktieren