Kanta Saito

Student at UH Manoa

I am studying for a B.S. in Computer Science in the Department of Information and Computer Sciences at the University of Hawaii. I expect to graduate in Spring, 2027.

INTERESTS: Artificial Intelligence · Machine Learning · Researching · Robotics · Computer Vision

CERTIFICATIONS:

• Machine Learning Specialization Certificate — DeepLearning.AI — May 2025
• 5-Day Gen AI Intensive with Google — Kaggle — April 2025
• Google AI Essentials — Google — February 2024

TECHNICAL SKILLS: Python, Java, C/C++, Windows/UNIX, SQL, NumPy, OpenCV, Pandas, Scikit-learn, TensorFlow

LANGUAGE SKILLS: Fluency in English and Japanese

Projects

Gamers Connect 2025-08-05

A full-stack web app that helps UH students find and connect with local gamers through matchmaking, sessions, and events.

Next.js TypeScript PostgreSQL Prisma Vercel

Linear Regression From Scratch 2025-03-01

A Python implementation of linear regression without using scikit-learn's built-in regression models. This project demonstrates the fundamental concepts of linear regression through a stock price prediction model for Apple Inc. (AAPL).

Linear Regression Python PCA

Neural Network From Scratch 2025-03-01

Implementation of a basic feedforward neural network from scratch using Python and the NumPy library.

Python NumPy MNIST Dataset Matplotlib

Gymenai 2024-03-15

I made a beginner gym AI agent for my 2025 GEN AI Intensive Captsone Project

Python Gemini SQL

CHANGE-HI Funded Research · 5-Minute Research Showcase

A 5-minute research showcase presentation on MammoClean, a project funded by CHANGE-HI (Center for Health Analytics, Notification, Guidance, and Education for Hawaii and the Pacific Islands).

School Projects

ICS 483 · Assignment 1

Implementation of the Hough Transform algorithm for detecting lines and shapes in images using edge detection and parameter space voting.

View PDF →

ICS 483 · Assignment 2

Computing planar homographies to overlay virtual objects onto real-world scenes, using feature matching and perspective transformations for augmented reality applications.

View PDF →

ICS 483 · Assignment 3

Reconstructing 3D geometry from 2D images using stereo vision, epipolar geometry, and triangulation to generate 3D point clouds from multiple camera views.

View PDF →

ICS 483 · Final Paper

Research paper on MammoClean, a preprocessing pipeline for mammography images that applies denoising and artifact removal to improve the accuracy of downstream machine learning models.

View PDF →

ECE 405

Exploring tokenization techniques and transformer architectures, covering subword tokenization methods, self-attention mechanisms, and the foundations of modern large language models.

View PDF →

Kanta Saito