Nathan Kiner, Machine Learning Developer in Paris, France
Nathan Kiner

Machine Learning Developer in Paris, France

Member since October 10, 2016
Nathan has been said to be a great team player and a rapid learner, at ease with technical complexity and business uncertainty. He has a master's degree in computer science and 4 years of experience building predictive models and data infrastructures in start-ups and as a freelancer. Nathan likes challenges and consistently makes sure to have his communication be at least as good as his development work.
Nathan is now available for hire

Portfolio

  • Google
    D3.js, Tableau, Python, BigQuery, SQL
  • Galvanize
    Scikit-learn, GraphLab, MongoDB, Flask, Spark, Python
  • Tractable
    Python

Experience

Location

Paris, France

Availability

Part-time

Preferred Environment

IPython, Atom, MacOS, Spark, SQL, Python

The most amazing...

...thing I did involved predicting the censorship rating from 100GB of movie subtitles using Spark/Doc2Vec, and leading the BI unit of a 1 million members startup.

Employment

  • Business Analyst

    2017 - PRESENT
    Google
    • Optimizing Google street car itineraries.
    • Building analytics interfaces and DW for process optimization.
    Technologies: D3.js, Tableau, Python, BigQuery, SQL
  • Data Science Fellow

    2016 - 2016
    Galvanize
    • Created a full-scale NLP pipeline with Spark and Doc2Vec to detect sensitive material in movie subtitles and classify the movies according to the MPAA censorship ratings.
    • Experimented with Frequentist and Bayesian frameworks for A/B Testing and applied to numerous use cases.
    • Detected fraud in a large dataset of transaction data and built a monitoring dashboard using Flask, pulling data from MongoDB and displaying them with matplotlib.
    • Analyzed, visualized, and predicted churn likelihood in transportation data from a ride-sharing app using Gradient Boosted Trees, SVMs, Random Forests, and ensembling techniques.
    • Built a recommendation engine for jokes based on like data from the Jester dataset using matrix factorization and collaborative filtering techniques.
    Technologies: Scikit-learn, GraphLab, MongoDB, Flask, Spark, Python
  • Data Science Consultant

    2016 - 2016
    Tractable
    • Researched Tractable.io and developed proprietary machine learning algorithms, with a focus on deep learning for computer vision.
    • Detected fraud in car insurance claims by implementing random walks on bipartite graphs. Python.
    • Predicted price of car parts by applying scalable collaborative filtering methods on car insurance audit data.
    Technologies: Python
  • Business Intelligence Manager

    2014 - 2015
    MONOQI
    • Led the business intelligence department and its business analyst team for a 1 million member startup.
    • Ensured data quality and information access by building dashboards and data pipelines from various sources and designing the data warehouse. Used SQL, AWS S3/Redshift/DP, and Ruby on Rails.
    • Accelerated daily newsletter generation time by a factor of 10 by initiating and supervising partial automation of the creation process.
    • Automated bidirectional data flows between the internal data warehouse and external marketing partners using Python.
    Technologies: Amazon Web Services (AWS), Python, Ruby on Rails (RoR), Ruby, D3.js, AWS Data Pipeline Service, Amazon S3 (AWS S3), Redshift, PostgreSQL
  • Junior Data Scientist

    2013 - 2014
    MONOQI
    • Identified 70% of future buyers by building a neural network model on early days activity. R, SQL.
    • Built and managed 100+ analytics dashboards and a custom interface for visualizing metrics. Used Ruby on Rails, NVD3.js, and SQL.
    • Increased conversion rate by 20% for members in disengagement phase through A/B Testing programs.
    Technologies: D3.js, PostgreSQL, Ruby on Rails (RoR), Ruby, R

Experience

  • Visualization of Multimedia Datasets
    https://github.com/Nathx/d3_cartography

    Prototyped an interactive interface for displaying large volumes of multimedia data (cartography, stream graphs, and video tapestry).

    Built using D3.js, Paper.js, and Raphael.js.

  • Parental Advisory Machine Learning

    A text classification and pattern detection algorithm on movie subtitles for censorship rating prediction and discovery of underlying features.

    Built using Python, Spark, Doc2Vec, Selenium, and Scrapy.

Skills

  • Languages

    Python, SQL, Ruby, R
  • Paradigms

    Data Science
  • Storage

    PostgreSQL, Redshift, MongoDB, Amazon S3 (AWS S3), AWS Data Pipeline Service
  • Other

    Machine Learning, Data Mining
  • Frameworks

    Apache Spark, Spark, Flask, GraphLab, Paper.js, Ruby on Rails (RoR)
  • Libraries/APIs

    Spark ML, Scikit-learn, NVD3, Raphaël, Matplotlib, D3.js
  • Tools

    Google Analytics, Atom, IPython, BigQuery, Tableau
  • Platforms

    Amazon EC2, Linux, MacOS, Amazon Web Services (AWS)

Education

  • Certificate in Data Science
    2016 - 2016
    Galvanize - (via online at http://www.galvanize.com/)
  • Engineer's/Master's Combined Degree in Computer Science
    2008 - 2012
    École Centrale Paris - Paris, France
  • Bachelor's Degree in Physics & Engineering
    2006 - 2008
    Lycée et Collège LAKANAL - Sceaux, France

To view more profiles

Join Toptal
Share it with others