Andrew Fief

Andrew Fief

About

Detail

Founder
Corvallis, Oregon, United States

Timeline


work
Job
school
Education
folder
Project

Résumé


Jobs verified_user 0% verified
  • Tabs
    Founder
    Tabs
    Mar 2026 - Current (5 months)
    Building a mobile app for debt tracking between friends Tabsapp.us
  • D
    Machine Learning Developer
    Dutch Bros Coffee Independent Project
    Sep 2025 - Apr 2026 (8 months)
    As a DB Barista - First: Did net sales forecasting for my franchise using upcoming local events, weather, and menu promotions. Then: HQ took interest in my project. Enabled me to build an ingredient parser that transforms order receipts -> individual ingredient volumes sold using real company data. Finally: Built a drink recommendation engine using pieces of my ingredient parser.
  • Dutch Bros Coffee
    Barista
    Dutch Bros Coffee
    Jun 2023 - Current (3 years 2 months)
Education verified_user 0% verified
  • Oregon State University
    Bachelor of Science - BS, Computer Science
    Oregon State University
    Jan 2022 - May 2026 (4 years 5 months)
    4th year computer science systems student at Oregon State University with a passion for software engineering and interest in machine learning. Oregon State University's Computer Science Systems degree is an accredited degree focused on fundamentals of Computer Science and Information Systems including programming, software engineering, discrete mathematics, computational theory, and hardware to open doors towards full-stack software development, embedded systems, system design, and other information systems.
Projects (professional or personal) verified_user 0% verified
  • E
    Energy Load Forecasting Web App
    May 2025 - Jun 2025 (2 months)
    A full-stack energy forecasting web application that predicts NYC's electricity demand 7 days ahead using machine learning and real-time weather data. The project demonstrates how historical power grid patterns can be combined with meteorological data to create accurate hourly load predictions, featuring automated data pipelines, model training, and a responsive dashboard for visualizing both weather forecasts and power demand predictions. Machine Learning Pipeline: - Implemented XGBoost regression model trained on 20+ years of NYISO grid data (2001-2025) - Engineered time-based features (hour, day, season) and temperature correlations for optimal performance - Achieved an average forecast accuracy of 97% per hourly prediction (mean absolut
  • D
    Digital Library - A SQL Database Project
    Apr 2025 - Jun 2025 (3 months)
    A collaborative web application for managing digital library content including books, authors, and genres, built with Node.js and MariaDB. The project was made in a small team, and demonstrates robust backend architecture using normalized relational database design, stored procedures, and a structured approach to data operations. Database Design & Logic: - Designed and implemented a relational schema fully normalized to Third Normal Form to eliminate redundancy and ensure data integrity. - Used a many-to-many intersection table to model relationships between books and genres. - Created stored procedures for all Create, Update, and Delete (CUD) operations to encapsulate logic and improve maintainability. Node.js Backend: - Developed a backen
  • S
    Solving the MNIST Dataset From Scratch in Python
    Mar 2025 - May 2025 (3 months)
    A full-stack digit classification web application powered by a custom deep neural network built entirely from scratch (without using machine learning frameworks like TensorFlow or PyTorch). The project demonstrates how a foundational neural network can be implemented using only NumPy, and it includes a Flask-powered backend and a JavaScript frontend to provide real time digit prediction for hand (mouse) drawn digits. The project is fully containerized and deployed with Docker and Docker Compose, with a container for both the front and back ends. Backend Neural Network: - Implemented a DNN with two hidden layers including sigmoid and softmax activations. - Designed and trained the model using raw CSV data from the MNIST dataset. - Achieved o
  • A
    Analyzing Walmart Sales Data to Understand Key Sales Drivers
    This project explores Walmart sales data from Kaggle to uncover key factors influencing sales performance. Using both supervised and unsupervised machine learning techniques, the analysis identifies patterns, trends, and relationships within the dataset. After data gathering and extensive data cleaning and preparation with pandas and NumPy, a structured quantitative dataset is created for analysis. For unsupervised learning, KMeans clustering from scikit-learn is applied to segment sales data into meaningful groups, providing insights into customer purchasing behavior. Supervised learning is then used with decision tree models to predict sales trends based on various influencing factors. Throughout the project, data visualization using matp
  • T
    TaskFlow: A Microservice-Based Task Management Application
    This task management application, built with Python’s Flask, is designed around a microservices architecture developed collaboratively by my team and me. It features modular services for language translation, email notifications, and data export, all communicating through RESTful APIs to ensure scalability and flexibility. The system supports core CRUD operations, including POST, DELETE, and PUT, enabling seamless task management. With optimized error handling, robust API request processing, and efficient JSON parsing, the platform delivers a reliable and scalable solution for distributed environments.
This is a community-created genome.