NLP Research
Language Modeling
Focused on Transformer-based machine translation, attention analysis, and evaluation design. Building from theory to implementation with reproducible experiments.
Portfolio - 2026
Alex H.
Applied Statistics & Mathematics
University of Toronto
01
About Me
Statistics is the language through which uncertainty becomes knowledge.
I'm a student at the University of Toronto pursuing an Applied Statistics Specialist and Applied Mathematics Major. My work sits at the boundary between rigorous statistical theory and practical machine learning, with a particular focus on Natural Language Processing.
I'm drawn to problems where mathematical structure meets real-world complexity — building models that are both theoretically grounded and genuinely useful. From Bayesian inference and probabilistic modeling to deep learning architectures, I enjoy the full spectrum from theory to implementation.
Outside of research, I care deeply about clean, reproducible code and the kind of collaborative science that moves the field forward.
Statistical Modeling
Applied Inference
Applied Bayesian thinking and statistical learning methods to practical prediction tasks, with emphasis on uncertainty interpretation and robust model diagnostics.
ML Systems
End-to-End Pipelines
Delivered data cleaning, feature engineering, model training, and validation workflows across coursework projects in Python, NumPy, scikit-learn, and PyTorch.
Visualization & Tools
Interactive Analytics
Built interactive dashboards and desktop tools, combining statistical storytelling with practical UI design in R Shiny, Plotly, Java Swing, and custom app workflows.
02
Education
2023.9 — Present
B.Sc. Applied Statistics Specialist
Applied Mathematics Major
University of Toronto, Faculty of Arts & Science
Specialist program in statistical theory, machine learning, and stochastic processes. Core courses: Statistical Learning, Bayesian Methods, Regression Analysis, Mathematical Statistics.
03
Projects
Project 01
Painting Classification Pipeline
End-to-end ML pipeline for art classification as part of the CSC311 challenge. Built data cleaning, feature engineering with TF-IDF, and trained Logistic Regression models achieving competitive accuracy with a numpy-only inference script for submission.
PythonNumPyscikit-learnTF-IDF
View Details →
Project 02
PaintApp — Interactive Drawing Tool
A feature-rich desktop painting application built from scratch with an intuitive UI for sketching and image editing. Supports multiple brush styles, color palettes, shape tools, and undo/redo history with a clean, responsive canvas experience.
JavaSwingOOPGUIMVC
View Code →
Project 03
Interactive Statistical Dashboard
R Shiny application with interactive ggplot2 and Plotly visualizations. Features adaptive label positioning for vertical reference lines across polynomial and KNN regression modes with dynamic axis scaling.
RShinyggplot2Plotly
View App →
Project 04
Transformer-based Machine Translation
From-scratch PyTorch implementation of the Transformer (Vaswani et al., 2017) trained on WMT14 en-fr and en-de without pretrained weights. Achieved 35.87 BLEU on en-fr with the Big model (209M params). Features SentencePiece BPE, BF16 mixed-precision training, beam search with checkpoint averaging, and published model weights on HuggingFace.
PyTorchTransformerNLPSentencePieceHuggingFace
View Repo →
Python
R
C
Java
PyTorch
scikit-learn
Hugging Face
NumPy/SciPy
Pandas
Docker
Git/GitHub
Linux
Matplotlib
04
Life & Interests
"The purpose of computing is insight, not numbers."
— Richard Hamming
Beyond mathematics and code, I care about understanding how things work at a deeper level — whether that is a statistical model, a piece of music, or a philosophical argument. Curiosity is the common thread behind my research and my daily life.
I am especially interested in the philosophy of statistics: what probability means, how we reason under uncertainty, and how assumptions shape conclusions. Those questions influence how I design models, evaluate evidence, and communicate results.
Music has been a core part of my life since childhood. I studied cello for years and played in school symphony settings, which taught me disciplined practice, collaboration, and sensitivity to structure and rhythm.
Outside academics, I enjoy badminton, long walks, and late-night debugging sessions that often turn into mini learning adventures.
05
Contact
Get In Touch
Contact
Live Decode