Close

Felix Fu

🏀👨‍💻🥾📚

Download Resume

About Me

I'm an undergraduate student pursuing an Bsc. CS Honours degree at the UBC. My previous summer work involves developing bioinformatic pipelines to process high dimensional sequencing data, using Probabilistic Programming Languages (Pyro/PyMC) for Bayesian modeling to infer clone population based on genomic alteration profiles (CNA, SNV) extracted from ctDNA. MCMC and VI are used to address the intractable integrals common in Bayesian inference. PPL encapuslates all the inference procedures and equations of various distributions which largely reduced the amount of work to implement a probabilistic model. A side project is about quality control for imaging mass cytometry (IMC) data using a Bayesian model. I completed my honours thesis about autoencoders on Pan-cancer ICGC dataset. You can find my thesis here. Any comments are welcomed! Currently, my work focuses on stochastic trajectories generation in various scenes using diffusion model. Our objective is generate trajectories that are controllable and realistic.

My favorite sport is basketball. I love to cook, go hiking, watch movies/series/NBA games in my spare time.

Experience

Summer Student - USRA

Roth lab, Molecular Oncology Department, BC Cancer Centre

  • Developing a Bayesian model for analyzing cfDNA (extracted from liquid biopsies) to estimate tumor fractions with graduate students. The model estimates the relative proportions of existing clones, and discovers emergences of novel clones based on copy-number profiles.
  • Applied the model to real world data from a cohort of 10 patients with matched primary tumour sequencing and longitudinal cfDNA sequencing.
  • Refined the graphic models and implemented them by PPL.
  • Deployed jobs on HPCs using Slurm and introduce parallelized workflows using Snakemake as well as making data analysis scalable and reproducible.

Undergraduate Teaching Assistant

UBC Department of Computer Science - CPSC210 (2 winter terms)

  • Design, development, and analysis of robust software components. Topics such as software design, computational models, data structures, debugging, and testing.
  • Received excellent evaluation scores and favorable feedback from 30+ students in preparedness and teaching quality.
  • Helped students to build and debug their projects in Java heuristically. Promoted the necessity of adhering design patterns during software construction.
  • Led 30-student labs both online and in-person and held customized recap or Q&A session. Relayed common problems and demands to professors.

Research Intern - CASIA

The State Key Laboratory of Management and Control for Complex Systems

  • Engaged in deep reinforcement learning (DRL) group meetings in different orientations. Gained a comprehensive understanding of DRL and its significant role in computer vision, autonomous driving and neural architecture search etc.
  • Learnt to construct computational graph and derived backpropagation method for calculating gradient. Evaluated the efficiency of several optimization methods (Momentum, Adam, AdaGrad) for SGD and regularization such as Dropout.

Lab Assistant

School of Automation at WHUT

  • Learnt about adaptive multi-task Probabilistic Movement Primitives in robot skill learning. Proofread and curated related documents.
  • Learnt about Gaussian Mixture Model and contributed to a paper that had been published fortunately.

Education

University of British Columbia

Sept 2019 - May 2023

Bachelor of Science in Honours Computer Science

Learn how to use and improve computers while exploring topics such as databases and operating systems, software engineering, security, web development, and numerical methodology. Computer Science gives me a thorough grounding in computer software design, and a broad choice of other studies in computing.


Rewarding courses taken:

  • "You might encounter plenty of oscillations in your life later on, hopefully only the mathematical types."
    -- Math 320 Real Variable
  • "If you ask for only one takeaway for this course, remember CACHE is everywhere"
    -- CPSC 313 Computer Hardware and OS
  • "The more we reduce ourselves to machines in the lower things, the more force we shall set free to use in the higher"
    -- CPSC 320 Algorithmic Analysis and Design
  • "It is all about ergodicity, reversibility and recurrence."
    -- Math 303 Intro to Stochastic Process

Hubei Wuchang Experimental High School

Sept 2015 - Sept 2018

Achievements & Awards

Work Learn International Undergraduate Research Awards

Department of Computer Science, Faculty of Science, UBC

The Work Learn International Undergraduate Research Awards are designed for international undergraduate students who are interested in research. This is equivalent to NSERC award for domestic students.
These awards subsidize professors to hire students to work on their research projects. International students who receive the awards gain workplace experience through an undergraduate research opportunity.

Trek Excellence Scholarship

Faculty of Science, UBC

Grant value: $4000
Trek Excellence Scholarships are offered every year to students in the top 5% of their undergraduate year, faculty, and school. Domestic students receive an award of $1,500 and international students receive an award of $4,000. Awards of $1,000 are also given to international students who are in the top 5% to 10%.

J Fred Muir Memorial Scholarship

Faculty of Science, UBC

Scholarships totalling $32,000 have been endowed through a bequest in memory of J. F. Muir by Lillian Muir. The awards are offered to students in the Faculty of Science on the recommendation of the Faculty.

Faculty of Science International Student Scholarship

Faculty of Science, UBC

Grant value: $10000
To recognize international students in the Faculty of Science who “demonstrate strong academic achievement, engagement in the Faculty, and the potential to make a scholarly contribution within their chosen field of study.”

Science Scholar/Dean's Honour List

Faculty of Science, UBC

Students promoted to second, third, or fourth year with a standing of 90% or better in the previous Winter Session will receive the notation "Science Scholar" on their records if they passed all courses completed and carried a course load of at least 27 percentage-graded credits.

Projects

Bayesian Model for Monitoring Cancer Progression Using Liquid Biopsies

A Bayesian Network for inferring tumour fraction and clonal prevalences from whole genome sequencing of cell-free DNA (cfDNA) based on genomic alterations (CNAs and SNVs) and Direct Library Preparation (DLP+) of a matched tissue biopsy.

View Project

University Student Center

An application that mimics UBC Student Service Center with HTML parser. This is an interactive platform where students could view schedule; edit courses; view grades and teaching assistants can create, grade assignments for all students

View Project

Amusement Park Management System

APMS concentrates on the domain of themed parks. The database of our project models the employment, logistics, and ticket services of the park. The whole project is built on Django framework and SQLite. We initiated with a comprehensive ER diagram respect to the staff and facilities in the park and normalized all the relation schemas in BCNF to reduce redundancy.

View Project

Insight Facade at UBC

A full-stack web application with front-end bootstrapped by Jumbotron. Devised Mocha unit test suite and the code passed eslint. Invoke the paradigm of asynchronous programming (i.e. properly handling promises). A back-end server implemented by Typescript to parse the designated queries about courses, buildings and rooms inside UBC. Followed the SOLID design principles and established final RESTful APIs.

View Project

Network gadgets suite

A Java GUI for dict server while connecting to it via Java Socket.
A fully functioned DNS server including resolving domain name, looking for the IP address by recursive approach and a command UI for email protocols such as SMTP/POP3.

View Project

Rendering using WebGL

A cute armadillo staring at a scalable/contractive coronavirus. Developed in WebGL and Three.js.

View Project

Identify Pulsar Stars utilizing KNN classification model

Pulsars are intensively magnetized, rapidly rotating neutron stars that emit periodic radio frequencies. I preprocessed the dataset HTRU2 and trained the KNN classifier through cross-validation following the pipeline. At last, I visualized the distribution of original data by ggplot2 and evaluated performance of classifier via confusion matrix.

View Project

Skills

Get in Touch