Search for a freelancer Post a project
Malt welcom

Welcome to Matteo's freelance profile!

Malt gives you access to the best freelancers for your projects. Contact Matteo to discuss your project or search for other freelancer profiles on Malt.

Matteo Guzzo

Senior Data Scientist

Can work in or around Berlin

  • 52.517
  • 13.3888
  • Suggested rate €833 / day
  • Experience 7+ years
Propose a project The project will begin once you accept Matteo's quote.

This freelancer is available full-time but hasn't confirmed their availability in over 7 days.

Propose a project The project will begin once you accept Matteo's quote.

Location and workplace preferences

Location
Berlin, Deutschland
Can work onsite in your office in
  • around Berlin and 50km

Preferences

Project length
  • Between 1-3 months
  • Between 3-6 months
  • ≥ 6 months

Verifications

Freelancer code of conduct signed
Read the Malt code of conduct

Verified email

Languages

  • Italian

    Native or bilingual

  • English

    Fluent

  • French

    Fluent

  • German

    Conversational

Categories

Skills (44)

Matteo in a few words

I build machine-learning products and assist my clients at all stages of development (end-to-end). I build complete Data Science pipelines, from data collection and filtering, to development and training of machine-learning models, ending with production deployment. This includes work needed to make stuff work "in real life" e.g. data engineering and software development.

Experience

Self employed

Digital & IT

Freelance Data Scientist  - As a freelancer

Berlin Area, Germany

April 2017 - Today (5 years5 years and 4 months)

I build machine-learning products and assist my clients at all stages of development. I can build a complete Data Science pipeline, from data collection and filtering, to development and training of machine-learning models, ending with production deployment. This includes work needed to make stuff work "in real life" e.g. data engineering and software development.
Recent projects:
- Author of Big Data Technologies course book for master program in Data Science and AI (with IUBH University)
- Data ingestion pipeline for drug discovery in cheminformatics (with Hotspot Therapeutics)
- Document discovery system for research archive (with Königsweg GmbH and Daimler AG)
- Question-answering system in the field of rail infrastructure law (with Cordian GmbH and Deutsche Bahn AG)
- Factorization machines of big data for user targeting in e-commerce (with Scout24 AG);
- Web application for performance marketing forecasting (with Pivigo LTD);
- Vehicle routing optimization for logistics (with Idalab GmbH);
- Network analysis of social media data with graph databases (with Unbound LTD));
- Time-series forecast of supplier payments in transportation (with Idalab GmbH);
- Data product for content marketers (Marketing trends prediction algorithm, ETL, integration of NLP models) (with Pivigo LTD and Waive LTD);
- Machine-learning models for education (unsupervised clustering of schools for disadvantaged students performance prediction, sponsor recommender system for under-performing schools) (with Pivigo LTD and DfE, UK);
- Machine-learning pipeline for purchase prediction in the house movers market (with Pivigo LTD).

AI Guild

Founding Member

Berlin Area, Germany

August 2019 - Today (3 years3 years)

Waive - Intelligent Content Marketing

Lead Data Scientist/Data Engineer (contractor)  - As a freelancer

London, United Kingdom

July 2017 - March 2018 (8 months)

Science to Data Science

Data Science Fellow  - As a freelancer

London, Greater London, Großbritannien

March 2017 - April 2017 (1 month)

S2DS is a five-week data science program for analytical PhD who are transitioning from academia to data science in the corporate world. The project of my team was with the UK-based firm TwentyCI. The goal was to implement a ranking system and predict future customers for furniture retailers. Overall results were: - Implemented machine learning pipeline in R for prediction of future customers. - Replaced hand-crafted reporting with automated pipeline for data munging and model training, ready for integration with Apteco FastStats cloud infrastructure (PMML output). - 4x improvement on success rate compared to pre-existing system (6.5% vs 1.6%). - Predicted month of purchase with 25% accuracy (vs. 9% random baseline).

Humboldt-Universität zu Berlin - Humboldt-Universität zu Berlin, Institut für Psychologie

Alexander von Humboldt Fellow (Wissenschaftlicher Mitarbeiter)

Berlin Area, Germany

January 2013 - December 2016 (3 years3 years and 11 months)

Main responsibilities: - Analytic development of quantum many-body methods for electronic structure calculations - Implementation of developed methods to python code - Numerical simulations in combination with existing Fortran code for distributed computing.

Ecole Polytechnique

Post-doctoral researcher

October 2012 - January 2013 (3 months)

Ecole Polytechnique

Agent CNRS

May 2009 - August 2009 (3 months)

Education