Frederick Chan
I'm a dependable software engineer who's passionate about writing maintainable software that developers and users alike love to work with. I follow the best-practices for your codebase, and carefully plan my code before I write it. I have a proven track record developing for diverse range of applications, including full stack web development, natural language processing, and machine learning. That's why projects I've worked on win awards and get published in cutting-edge scientific journals like Science!
Skills
- Python
- JavaScript
- TypeScript
- Java
- C++
- Lua
- SQL
- Ruby
- HTML5
- CSS
Languages:
- Node.js
- React
- React Native
- Pandas
- TensorFlow
- PyTorch
- Ruby on Rails
- Git
- Github
- Linux
- QGIS
- Docker
Technologies:
- Team leadership
- stakeholder communication
- research
- writing
Soft skills:
Professional experience
-
Conzensus, LLC
Software Engineer -
- Led a team of 5 in creating the architecture and software requirements for a real-time, cross-platform mobile application starting from the concept stage
- Implemented both the server and client side of a full-stack application with React Native, Node.js, and JavaScript
- Conducted user research and collaborated with a team of 5 to design and prototype app UX designs
-
Institute for Protein Design, David Baker Lab
Research Software Engineer and Web Developer -
- Co-author of Robust deep learning based protein sequence design using ProteinMPNN, a PyTorch neural network with 19.5% better sequence recovery than DeepMind's AlphaFold
- Adapted ProteinMPNN for protein backbone structure prediction for greater sequence diversity in novel proteins using an approach inspired by diffusion neural networks
- Developed the Foldit website redesign with Ruby on Rails and JavaScript
- Conducted feasibility studies for integrating a modern HTML5-based UI into Foldit, written in C++, to improve code maintainability
- Created and integrated my own webpage designs into existing UX based on stakeholder feedback
-
Newbook Digital Texts
Full Stack Developer Intern -
- Designed and developed Nile Travelogues, a full stack web app using Node.js, SQLite, and JavaScript
- Reduced manual work up to 90% when converting print information to a digital database using Python and Wolfram Language
- Developed an application to automatically annotate historical documents with information about text entities from Wikidata using NLP techniques
Education
BSc Informatics (Data Science) & BA Linguistics (double degree, with honors)
University of Washington, Seattle -
- Leadership: 1 yr. President and 2 yrs. Officer of the Conlang Club at UW
- iSchool Informatics Innovation Award winner
- Graduated Magna Cum Laude in Informatics and Cum Laude in Linguistics
- Washington State Opportunity Scholarship recipient
Talks and publications
- 2022, Science: Robust deep learning-based protein sequence design using ProteinMPNN
- 2021, eScience Institute: They Came to Egypt: Databases and Digital Editions of 19th Century Nile Travelogues
Projects
(See my Github profile at github.com/fechan for more)
- Grow Online (D3, Node, JavaScript): Web implementation of a board game I invented called Grow that's playable on a web browser.
- Fredbox (React, Node, JavaScript): Multiplayer game where players compete to answer a series of brain games as quickly as possible inspired by Jackbox.
- Trilangle (D3, JavaScript, Formal Language): Constructed language using tessellating triangles to encode syntax and represent meaning. Has a sentence renderer in D3 and JavaScript.
- Mathematikka (Java, Mathematica): Bukkit server plugin that uses the Mathematica kernel to query WolframAlpha from within Minecraft.
- Minecraft Geological Survey (Python): Tools for mapping and analyzing Minecraft maps in Geographic Information Systems (GIS) software.
- Legislature Simulator (JavaScript): Online zero-player game that simulates a fictional country's legislative sessions.
Volunteering
- Sep 2022 - present, Seattle Conlang Club, Club President
- Jun 2016 - Jan 2017, Seattle Public Library, Chess Club Volunteer