Edwin van der Helm, Physics Simulations Developer in Calgary, AB, Canada
Edwin van der Helm

Physics Simulations Developer in Calgary, AB, Canada

Member since May 21, 2019
Edwin has been at the heart of the deep learning wave since it gained industry popularity. Before that, he has worked as a full-stack Java developer and completed a PhD in computational astrophysics. He is a social guy who communicates well and excels at working with teams, learning from others' domain knowledge and sharing his technical knowledge freely.
Edwin is now available for hire

Portfolio

Experience

Location

Calgary, AB, Canada

Availability

Part-time

Preferred Environment

Vim Text Editor, Unix, Git, Pandas, TensorFlow, Python

The most amazing...

...product I've written is a user-friendly library for distributed automated hyperparameter optimization.

Employment

  • Solution Architect | Neural Network Engineer

    2015 - PRESENT
    Minds.ai
    • Created an internal toolset to augment TensorFlow and manage the neural network training runs for enterprise projects.
    • Constructed an advanced and easy-to-use, distributed automated hyperparameter optimization library.
    • Designed and built a neural network-based named-entity recognition system for medical literature.
    • Led a distributed data-science team and created the main neural network for text search based on semantic meaning in addition to keyword matches.
    • Trained a classification neural network for database tables.
    • Instructed and supervised other engineers in the expected use of source control, code style, documentation, and continuous development tools.
    • Supported the marketing and BizDev team by creating technical presentations, project plans, demos, and meeting with potential customers.
    Technologies: SpaCy, TensorFlow, Python
  • PhD Candidate

    2012 - 2016
    Leiden Observatory
    • Extended the AMUSE framework used for multi-physics astrophysical simulations at large hardware scales.
    • Published multiple scientific papers on hydrodynamics, star formation, black holes, and stellar wind.
    • Wrote a Python library for combining hydrodynamics, N-Body, and stellar evolution simulation codes to simulate stellar wind in embedded star clusters.
    • Taught multiple undergraduate- and graduate-level courses on Python programming and computational astrophysics.
    • Supervised a master's thesis research project.
    Technologies: Python
  • Java Developer

    2010 - 2012
    Atos
    • Worked in a team that created a civil registration service for the Dutch government.
    • Designed and extended domain-specific languages using Xtext to speed up both database design and front-end development.
    • Extended a functionality of FileNet for mail and document processing in Insurance companies.
    • Created and safeguarded software development standards for a distributed team in the Netherlands and India.
    • Performed a software quality audit for a critical application at a major international bank.
    Technologies: Xtext, Java

Experience

  • Named Entity Recognition

    As a proof of concept for a product, I prepared a large dataset using only publicly available data and trained a neural network for named-entity recognition in medical publications. Because there was no dedicated training data, I had to very carefully combine and augment the data that was available and adjust the neural network architecture to use it effectively.

    In a few months, I was able to get results that could compete with models trained on proprietary data that had cost other teams years to compile.

  • Automated Hyperparameter Optimization

    I have designed and implemented an easy-to-use distributed automated hyperparameter optimization library that combines the strengths of different optimization techniques, including HyperBand, Tree of Parzen, regression trees, and random forests.

  • Semantic Search

    I led a large distributed team to find and prepare training data on financial news articles. For those articles, I also designed and trained a neural network to find relevant articles based on semantic text content.

  • Database Table Classification

    I designed and trained a neural network for classifying database tables into several thousand custom classes. Since many of these tables were very large, the data pipeline had to efficiently process numerous terabytes of training data.

  • TensorFlow Toolset

    Since TensorFlow was first made publicly available, I have been the lead architect and developer of a library for streamlining neural network development for enterprise customers; including training run management, visualization, data processing, augmentation, and hyperparameter optimization.

Skills

  • Languages

    Python, Java, Java 6
  • Libraries/APIs

    TensorFlow, Pandas, SpaCy
  • Paradigms

    Data Science, Object-oriented Programming (OOP), Distributed Computing
  • Platforms

    Jupyter Notebook, Linux, Unix
  • Other

    Data, Natural Language Processing (NLP), Deep Neural Networks, Data Visualization, Physics Simulations, Computational Physics, Remote Team Leadership, Convolutional Neural Networks, Machine Learning, Algorithm Visualization, Information Visualization, Visualization, University Teaching, GPU Computing, Team Leadership, Mathematics, Recurrent Neural Networks, Web UI, Deep Reinforcement Learning
  • Tools

    Vim Text Editor, Git
  • Frameworks

    Xtext

Education

  • PhD Degree in Astrophysics
    2012 - 2016
    Leiden Observatory - Leiden, Netherlands
  • Master's Degree in Astrophysics
    2007 - 2009
    Utrecht University - Utrecht, Netherlands
  • Bachelor's Degree in Science
    2004 - 2007
    University College Utrecht - Utrecht, Netherlands

Certifications

  • Certified Java Programmer
    FEBRUARY 2011 - PRESENT
    Oracle

To view more profiles

Join Toptal
Share it with others