Seraphina Goldfarb-Tarrant, Machine Learning Developer in Santa Rosa, CA, United States
Seraphina Goldfarb-Tarrant

Machine Learning Developer in Santa Rosa, CA, United States

Member since September 23, 2019
Seraphina has years of experience as a machine learning engineer and NLP researcher in Python. She's developed state-of-the-art ML models that have been published in major conferences. Seraphina has also worked as a product manager at Google, where she led teams around the world that grew from a handful of people to multibillion-dollar businesses. She loves to solve new NLP challenges and works smoothly across different cultures and timezones.
Seraphina is now available for hire



  • Python 8 years
  • Machine Learning 5 years
  • Natural Language Understanding 3 years
  • Natural Language Processing (NLP) 3 years
  • Neural Networks 2 years
  • PyTorch 1 year


Santa Rosa, CA, United States



Preferred Environment

GitHub, Linux, Anaconda

The most amazing...

...system I've coded was an ensemble of neural networks that could collaboratively write stories with the user, and focus on specific attributes of good writing.


  • Machine Learning Engineer

    2019 - PRESENT
    Bayes Centre for Data Science
    • Developed a Python pipeline to scrape the web for scholarly articles on African livestock disease and extract necessary data for researchers to use—resulting in making a more than three-month-long human process into an overnight job.
    • Wrote a classifier to select scholarly articles as relevant vs. not relevant—with 85% accuracy over 12 Subsaharan regions.
    • Wrote an information extraction system to extract data fields from articles for use in large scale data analysis - processing hundreds of articles at a time.
    • Created an API and user interface for veterinary researchers to interact with the system after project completion.
    • Ran a research experiment to analyze the different options for information extraction and balance generalisability vs. training data requirements.
    Technologies: Pytorch, Scikit-learn, Neural Networks, Information Extraction
  • Research Engineer

    2018 - 2019
    Information Sciences Institute
    • Created an interactive demo where humans can interface with neural networks to cowrite stories together.
    • Served as the primary engineer for natural language generation and drove the research agenda.
    • Ran evaluations and experiments for neural networks and wrote and published the results.
    • Supervised interns and gave company-wide seminars.
    • Consulted with other teams on how to incorporate neural generation into their systems.
    Technologies: Pytorch, Python
  • Product Manager

    2013 - 2015
    • Launched an adtech product in new and emerging markets, focusing on Australia/New Zealand, Japan, and China; where I increased local partner integrations by 300%, leading to revenue growth of 212% YoY ($11M to $34 million).
    • Established localization efforts and language detection and categorization for ad products.
    • Launched A/B testing for our ML algorithm controlling bidding on ad space.
    • Changed the method of integration with Yahoo! properties, eliminating the primary source of on-call incidents.
    Technologies: SQL, Balsamiq, Sketch
  • Technical Account Manager

    2011 - 2013
    • Implemented custom solutions in Python/Django for the top 40% of Google adtech clients.
    • Provided on-call for instant resolution of adtech incidents.
    • Oversaw and was in charge of running QA for new releases.
    • Worked on a team that was migrating a tech stack from Amazon to Google infrastructure.
    • Developed materials and gave client training sessions on the tech stack and how best to utilize it.
    Technologies: Python, Django, C++
  • Digital Mapping Assistant

    2010 - 2010
    Experiential Technologies Center at UCLA
    • Developed 3D reconstructions of ancient Rome throughout time, to integrate a time dimension into Google Earth and assist in historical research.
    Technologies: KML, Google Earth API
  • Tutor

    2008 - 2010
    Academics in the Commons at UCLA
    • Taught small tutoring courses for undergraduate students in various fields, including C++, Ancient Greek, Astronomy, and English Composition.
    Technologies: N/A


  • Interactive Neural Storytelling (Development)

    A web application that allows any user to co-write stories with a state-of-the-art neural network that I developed, which tells commonsense stories. Users can either provide only a title (auto mode) or co-write a story entirely (interactive mode).


  • Languages

    Python, C++, SQL, Bash
  • Libraries/APIs

    PyTorch,, Twitter API, Reddit API, NumPy, Sklearn
  • Other

    Machine Learning, Neural Networks, Natural Language Processing (NLP), Natural Language Understanding
  • Platforms

    Amazon Alexa, Anaconda, AWS EC2
  • Frameworks

  • Tools

    Kaldi, Balsamiq, Sketch, Git, Jira
  • Storage

    AWS DynamoDB


  • Master of Science degree in Natural Language Processing
    2017 - 2019
    University of Washington - Seattle, WA, USA
  • Bachelor of Arts degree in Ancient Greek
    2006 - 2010
    University of California Los Angeles - Los Angeles, CA, USA

To view more profiles

Join Toptal
Share it with others