Data Scientist (f/m/d) wanted for Berlin!

Firmenname für PREMIUM-Mitglieder sichtbar

  • Oktober 2021
  • September 2022
  • D-Großraum Berlin | D-Berlin
  • auf Anfrage
  • Remote
  • 23.09.2021
  • 2755

Projekt Insights

seit wann aktiv?
Projektansichten
Bewerbungen

Projektbeschreibung

For an international customer we are currently looking for Data Scientist (f/m/d).

Your task will be:
• Implement an unsupervised deep learning based anomaly detection approach for imbalanced data
• Document the training techniques employed for the solution
• Implement the code in Python 3.7+ and either Pytorch or Keras (Tensorflow 2.0+) as the Deep Learning framework
• Ensure the code is available and reproducible on a GPU-enabled Linux machine
• Ensure all data, code used, workflow (including pre-processing and featureextraction), training and hyperparameters are tracked for every reported result
• Document the limitations of the approaches and recommend constraints to improve the performance of the Anomaly Detection algorithms
• Cooperate with the team through weekly 30-60 minute calls to discuss hypotheses being tested, findings and blockers or questions for domain experts

Following Know-how is required:
• Proven experience in Time Series analysis with Deep Learning approaches applied to industrial datasets
• Experience in other Machine Learning methods for time series analysis and applied statistics
• Comfortable with Pandas, Scikit-Learn, Numpy and either Pytorch, Tensorflow or another common DL framework
• Some experience with semi-supervised and self-supervised methods on structured data
• Some experience in NLP with non-academic datasets
• Some experience with at least one Python API framework
• Excellent communicator English
• Some experience with Git, Docker and experience scaling ETL pipelines

Order type: Contract
location: Berlin/remote
Start: 1.10.2021
Duration: 30.09.2022 ( with option of extension)

If you are interested, please tell us about your hourly rate and your availability.
We are looking forward to your application in an MS-Word-readable format quoting the reference-number 2755.

Kontaktdaten

Als registriertes Mitglied von freelance.de können Sie sich direkt auf dieses Projekt bewerben.

Sie suchen Freelancer?

Schreiben Sie Ihr Projekt aus und erhalten Sie noch heute passende Angebote.

Jetzt Projekt erstellen