PhD researcher · University of Waterloo

houfu@research:~$ simulate --world --one-system-at-a-time

I build small, working worlds.

I’m Houfu Chen — a digital twin researcher and engineer connecting mechanistic models, noisy sensor data, and machine learning to make complex systems tangible, testable, and understandable.

Explore selected work Start a conversation ↗
  • Python
  • PyTorch
  • COBRApy
  • SBML
SCROLL

01 / PROFILE

Engineer by training. Builder by instinct.

I work where models meet reality: noisy signals, heterogeneous environments, and systems that need to stay useful outside a notebook.

I’m a PhD researcher and engineer with one practical obsession: putting a working version of the world in my hands, then using it to ask better questions.

  • 01 Building living digital twins for heterogeneous environments.
  • 02 Integrating machine learning, mechanistic simulation, and measurement audits into reproducible workflows.
  • 03 Turning theoretical models into working systems that can be tested against reality.
~/about_me $ ls

Education

  • MEng in Electrical and Computer Engineering University of Toronto (Graduated March 2025)
  • BASc in Nanotechnology Engineering University of Waterloo (Graduated June 2023, Honours) President’s Scholarship of Distinction · President’s Research Award
  • Featured alumni in the Data Analytics and Machine Learning Emphasis at UofT. Read more here.

    Verbatim, from the UofT program page

    With the rapid growth of machine learning and the expansion of companies, machine learning and big data have become essential tool sets for ensuring continuous service availability and enhancing customer experiences. Mastering these areas enables me to drive data-driven strategies and strengthens my job prospects in today's competitive landscape.

    Houfu Chen, ECE MEng student read the full page
  • Credentials Machine Learning Specialization — Stanford University & DeepLearning.AI (2025) · SQL for Data Science — University of California, Davis (2025) · Algorithms and Data Structures Certificate — Stanford Online (2025)

Industry Experience

  • Ford Motor - SYNC & TCU Network Performance
  • Manulife - QA automation & cross-browser testing
  • Teranet - QA Automation
  • Imagine Communications - Quality Assurance

Teaching & Research

  • Teaching Assistant Machine Learning & Python at University of Toronto
  • Research Assistant Polymer Semiconductors & Organic Thin-Film Electronics at University of Waterloo

Project Highlights

  • Living Greenhouse Digital Twin
  • Smart Evacuation Maps
  • S.W.I.M (Sponsored by Prof. Steve Mann)
  • Browser Tools & Game of Life

Greenhouse-Level Digital Twin Framework

> Status: Active PhD Research · Prototype Integration

  • Building a dynamic flux-balance model of greenhouse tomato metabolism, working toward a living digital twin of a heterogeneous greenhouse.
  • The Core Architecture: Bridging the gap between mechanistic biological predictions and noisy real-world sensor data.
  • Mathematical Frameworks: An hourly Python/COBRApy workflow built on SBML models, sequential linear programs and parsimonious FBA, coupling a vegetative source network to a fruit model through nine phloem and xylem transport pools.
  • Validation Evidence: Model-to-measurement validation using 45 traceable tomato and leaf mass samples and a 23-day canopy-height series, which identified dry mass, fruit age, diameter, and Brix as high-value validation measurements.
  • Tech Stack: Python, COBRApy, SBML, Linear Programming, NumPy, SciPy
[🔒 Private research repository · active work]

System Capabilities

  • [SYS] Languages: Python, C++, JavaScript, Java, C#, R, SQL
  • [ML] Machine Learning: PyTorch, TensorFlow, Scikit-learn, PCA, Regression, Classification, Clustering, DeepFace, MediaPipe
  • [DAT] Data & Modeling: Pandas, NumPy, SciPy, COBRApy, SBML, dynamic FBA, pFBA, FVA, Linear Programming, Sensitivity Analysis, Parameter Calibration, Data Visualization
  • [CFG] Frameworks & Tools: React, Node.js, Git, Docker, Linux/Ubuntu, AWS Lambda, API Gateway, REST APIs, Azure, Selenium

02 / EXPERIENCE

Research is the thread. Range is the toolkit.

My current PhD work leads the story; teaching, product engineering, and consulting give me the range to carry ideas into real systems.

~/experience $ ls -l

Management / Consulting

Complaints Management & Internal Consulting (Intern) @ FAW-Volkswagen

Dispute Resolution, Risk Prevention

May 2025 – July 2025

  • Handled high-risk complaints escalated by frontline agents.
  • Assessed solution fairness and determined fault through cross-functional coordination.
  • Built and maintained QBI Kanban boards to track complaint progress.

CAN Tech Industry

WiFi Software Testing Engineer (Intern) @ Ford Motor Company

Ubuntu, iperf3, TCP/IP, Selenium

May 2022 – August 2022

QA-Developer (Co-op) @ Manulife

JavaScript, React, WebDriver IO

January 2020 – April 2020

QA Automation Specialist (Co-op) @ Teranet CMS

C#, Selenium

September 2020 – December 2020

Teaching Experience

APS1070 (Foundations of Data Analytics and ML) @ UofT

  • Graded 100+ projects and exams per term
  • Led 10+ Tutorials per term

CSC108 (Introduction to Computer Programming) @ UofT

  • Ran tutorials and office hours on Python fundamentals, debugging, and code style
  • Assessed projects and exams for courses serving up to 500+ students per term

MIE370 (Introduction to ML) @ UofT

  • Created 2 Projects, 2 Tutorials, 2 Exam Questions

Research Experience

PhD Researcher @ University of Waterloo

Living Digital Twins, dFBA, State Estimation

Current

  • Developing a living greenhouse twin that connects mechanistic biological models with noisy, heterogeneous sensor data.
  • Building a dynamic flux-balance model that couples a VYTOP-derived vegetative source network to a fruit model through nine phloem and xylem transport pools.
  • Implemented an hourly Python/COBRApy simulation workflow using SBML models, sequential linear programs, parsimonious FBA, and development-dependent fruit objectives.
  • Supervised by Prof. Nasser Mohieddin Abukhdeir and Prof. Christian Euler in Chemical Engineering.

Research Assistant @ Prof. Yuning Li's Lab, UWaterloo

Chromatography, UV-Vis Spectroscopy, Polymer Semiconductors

May 2019 – August 2019

  • Performed chromatography and UV-vis spectroscopy for polymer semiconductor and organic thin-film research.

Research Contributor @ Jilin Agricultural Science and Technology University

Python, PCA, Spectroscopy, Field-Trial Data

2025

  • Analyzed multi-year field-trial data spanning soil chemistry, fluorescence spectroscopy, FTIR, thermogravimetry, and crop yield using Python and principal component analysis.
  • Integrated multidimensional measurements into treatment-ranking analyses and supported manuscript structure, statistical interpretation, and technical editing.

03 / SELECTED WORK

Proof lives in the prototype.

Research systems and visual experiments share the same instinct: build a working method, test it against reality, and keep the evidence visible.

FEATURED SIDE PROJECT · REQUESTS OPEN

Teaching a model to paint light—not a texture filter.

Monet Light & Strokes is a benchmark-driven visual skill built through repeated image critique. It locks the source scene, lets color mix optically, and changes the brush vocabulary with the physical state of water, sky, foliage, stone, and shadow.

  • 01Scene facts stay fixed
  • 02Light carries the form
  • 03Benchmarks reject pretty mistakes
Submit your photos ↗ Send special requests ↗ Explore the open skill ↗ Open the gold standard ↗

Upload 1–3 photos through the private Dropbox request, then email what must stay unchanged, your special requests, intended crop, and sharing permission. Requests are handled personally—the website does not store uploads or call an image API.

Original photograph of a seabird touching sunset water ORIGINAL
Approved broken-color painting of the same seabird and sunset water BENCHMARK
01 / PHOTO → PAINTThe scene stays. The water learns to move.
Original photograph of cacti overlooking a town at dusk ORIGINAL
Approved painting of the same cacti and town beneath a separated-color dusk sky BENCHMARK
02 / LIGHTColor makes dusk luminous
Original night photograph of two figures beside an illuminated angular house ORIGINAL
Approved nocturne painting of the same figures, staircase, and illuminated house BENCHMARK
03 / ATMOSPHEREChaos holds the structure
~/selected-systems $ inspect

FLAGSHIP RESEARCH · UNIVERSITY OF WATERLOO

A living digital twin for heterogeneous greenhouses

> Active PhD research · prototype integration

  • ProblemGreenhouses are spatially uneven, partially observed, and full of noisy sensor signals.
  • MethodCouple a vegetative source network to a fruit model through phloem and xylem transport pools, simulated hourly with dynamic flux-balance analysis.
  • Current buildA running simulation plus evidence and equation audits that separate measured, literature-transferred, fitted and chosen parameters.
  • EvidenceModel-to-measurement validation with 45 traceable tomato and leaf mass samples and a 23-day canopy-height series flagged dry mass, fruit age, diameter, and Brix as high-value validation measurements.
  • GoalA twin that keeps updating as the real greenhouse changes — not a one-time simulation.

Supervised by Prof. Nasser Mohieddin Abukhdeir and Prof. Christian Euler · Chemical Engineering

ROUTING SYSTEM · OPEN NOTEBOOK

Building Smart Evacuation Maps

  • ProblemShortest-path routing can create unsafe crowd concentration during majority evacuations.
  • MethodCombine Dijkstra routing, ant-colony-inspired optimization, and LLM-assisted decisions.
  • EvidenceOutperformed simple Dijkstra in the tested majority-evacuation scenarios.
GitHub ↗ Build note ↗

PHYSICAL COMPUTING · LIGHT

S.W.I.M (Sequential Wave Imprinting Machine)

  • IdeaMove a one-dimensional light source through space to imprint a two-dimensional image.
  • BuildArduino, addressable LEDs, and time-lapse photography turn digital pixels into a physical trace.
  • ContextHardware project sponsored by Prof. Steve Mann.
GitHub ↗
Sequential Wave Imprinting Machine light pattern S.W.I.M. hardware experiment in motion

~/experiments $ ls --compact

Older experiments stay available without competing with the research story.

Multimodal emotion analysisDeepFace · MediaPipe · Ollama ↗ Personal finance trackerLocal-first · OFX/QFX → Autonomous retrieval robotLEGO Mindstorms · 3rd place ↗ Matrix calculatorPython · GUI ↗

04 / LIVE LAB

Useful tools, built in public.

A growing shelf of focused browser tools: read, reflect, track, play, and leave a signal.

~/live-lab $ ls --available

SWIPE TO BROWSE →

A neon cellular butterfly from Conway’s Game of Life FLAGSHIP · LIVE
[##]
game_of_life.exe
Draw patterns, watch them evolve, and turn them into music. Explore a sandbox, 11 leaderboard challenges, and a Music Box sequencer.
flagship sandbox challenges music
JS · PWA · leaderboard → play ↗
An old fisherman with a golden fish rising on his line, beside the game’s verdict — 妙手, a brilliancy NEW · LIVE
[⬡]
gomoku.exe
Five in a row against a minimax AI, a friend online, or itself. An old fisherman watches from the bank and reads every move — a fish on the line for a good one, an old tin for a blunder — then reviews the whole game at the end and replays the move that decided it.
the guru minimax · α-β online game review
JS · canvas · bilingual → play ↗
[📖]
reading_room.app
A private bilingual reader with local OCR, PDF and text import, line-by-line translation, paragraph notes, and portable backups.
local-first OCR FR → EN
Browser · private → open
[✎]
journal_intime.app
A calm space for small reflections and daily signals, designed to make writing feel lighter and more immediate.
reflection daily weather-aware
Journal · browser → write
[$]
fin_tracker.app
A teachable personal finance tracker for categorizing transactions, seeing patterns, and keeping private data under your control.
insights categories privacy
Data · local → track
[✉]
message_box.app
A low-friction way to leave a private note, question, or hello without needing to open an email client.
direct private no login
Inbox · async → leave a note
[◎]
attachment_quiz.app
An 18-question attachment-style explorer with a clear Anxiety × Avoidance visualization, shareable results, and compare mode.
visualization shareable compare
18 questions → explore

05 / NOTES & LIFE

There’s more than one mode.

Notes from the PhD journey, private-by-design experiments, and ideas I’m still turning over.

~/personal_logs $ ls -a

Interdisciplinary PhD Journey

Status: Current PhD @ University of Waterloo

  • Research Focus: Digital Twins & Systems Architecture
  • Department: Chemical Engineering
  • Supervisors: Prof. Nasser Mohieddin Abukhdeir & Prof. Christian Euler

Knowledge Base

A collection of my thoughts, articles, and book notes.

  • Computer Vision & AI: Notes from recent deep dives.
  • Algorithms: Problem-solving patterns and optimizations.
  • How to Be an Imperfectionist: Key takeaways on productivity.

Salle de lecture — bilingual reading room

A private reader for books I’m working through in French: import a passage, interleave English line by line, and attach notes to any paragraph.

  • Local OCR extracts highlighted text from screenshots
  • PDF and text import stay in the browser
  • Paragraph notes, markdown export, and JSON backup
→ Enter the reading room

~/life $ ./game_life --launch

Game Life

9 games · 3 co-op

A whole arcade with its own home — cellular automata, AI board games, a torch-lit roguelike maze, a plant-metabolism farming sim, and real-time co-op.

  • game_of_life.exe
  • gomoku.exe
  • labyrinth.exe
  • grow_a_tomato.exe
  • tomatoswipe.exe
  • sudoku.exe
  • keep_up.exe
  • lander.exe
  • tether.exe
Life is a game — enjoy every piece of it. Don’t be afraid. Don’t rush.
→ Enter the arcade ↗

Attachment Style Quiz

Discover your attachment style — Secure, Anxious, Avoidant, or Fearful — through 18 questions inspired by the ECR-R framework from “Attached” by Levine & Heller.

  • 2D scatter plot visualization on Anxiety × Avoidance axes
  • Shareable result links
  • Compare mode: plot two people on the same chart
→ Take the quiz

06 / CONTACT

Let’s make something useful.

I’m open to research collaborations, technical conversations, and ambitious systems problems — especially where modeling, sensing, and machine learning overlap.

Email Houfu