Suzen Fylke

Suzen Fylke

Machine Learning Engineer


I create products that simplify tedious processes in machine learning model development and deployment. My work enables researchers and engineers to iterate faster, transition from prototype to production more smoothly, and retrain their models automatically and reproducibly.

Work Experience

2024 — Now
Gothenbug, Sweden

Bringing efficient deep learning to embedded systems.

2020 — 2023
Machine Learning Engineer at Twitter
New York, NY, USA

Designed, built, and maintained ML pipeline infrastructure for teams operating at scale.

2019 — 2019
Software Engineering Intern at Twitter
New York, NY, USA

Prototyped a solution to streamline the new user experience of ML Workflows, an internal tool designed to automate, schedule, and share machine learning pipelines.

2018 — 2018
Software Engineering Intern at Twitter
San Francisco, CA, USA

Implemented some of the top ten feature requests for Twitter’s ad content moderation platform, increasing reviewer productivity and reducing unsafe ad impressions.


2016 — 2020
New York, NY, USA

Selected Coursework: Open Source Software Development, Language Technology, Supervised Research: Computational Models of Syntax Acquisition, Supervised Research: Using Kernel Methods and Model Selection to Predict Preterm Birth

2011 — 2015
New York, NY, USA

Selected Coursework: Computing in Context – Digital Humanities, Languages of America, Advanced Italian I, Intermediate Latin I, Advanced Workshop in Translation, Rapid Reading and Translation in Italian


2022 — Now

Model Card Toolkit is a library that that streamlines and automates the generation of model cards.

2020 — Now

TensorFlow Extended (TFX) is an end-to-end platform for deploying production ML pipelines.

Side Projects


Walt is a minimal documentation theme based on writ.css. It features classless styles for semantic HTML and supports light and dark mode based on system settings.


Terminal Block is a WordPress block plugin for displaying rich text in an animated terminal window. It's compatible with WordPress 5.9 or higher and is available in the Block Directory.


I once read a Tweet that said, "you can make any Wikipedia article dystopian by changing it to the past tense." Dystopian Wikipedia sounded like a fun idea, so I built Dystopedia as a proof-of-concept.



I explored the relationship between lemma length and CEFR level for lemmas in the Swedish Kelly list dataset. Here's what I learned!


The ability to visually inspect TFX artifacts with InteractiveContext is invaluable for debugging. Learn how to visualize custom TFX artifacts with InteractiveContext in this tutorial.


My reflections on learning how to code by printing "Hello, World!". Written two years after I started programming.