TRUNG BUI

adpm57@r.postjobfree.com

Github: github.com/TrungBui59 Linkedin: linkedin.com/in/trungbui2002 Education

Purdue University West Lafayette, Indiana

BS in Computer Science Honor. GPA: 3.77 Expected Graduation - August, 2024 Dean’s List August 2020

Semester Honor August 2020

Work Experience

Part-Time Machine Learning Engineer Intern ProtonX AI July 2021 - Present

- Created a full pipeline with testing for several computer vison architectures including VGG19, ResNet, MobileNet, and AlexNet, using TensorFlow 2.4.

- Optimized the architecture to achieve more than 90% accuracy and 4% faster than the original implementation

Undergraduate Researcher Duality Lab June 2021 - August 2021

- Discovered the bottleneck in the current engineering workflow of machine learning projects.

- Collected 300 reproducibility levels, bug manifestations, and fix patterns randomly using Github API.

- Analyzed the correlation between reproducibility level and fix patterns using Pandas and Matplotlib TensorFlow Model Developer GoogleML x Purdue August 2021 - Present

- Led and trained 10 undergraduate students to develop YOLO family models paper using ETL pipeline into deployable open-source models for TensorFlow’s Github repository, which has over 2900 forks on Github

- Achieved the same state-of-the-art result as the author’s original implementation. Projects

Cryptocurrency Trading Bot Pandas, NumPy, Sci-kit-learn, TensorFlow

- Developed a deep reinforcement learning trading bot using TensorFlow. The project generated a profit of 112% with $1000 investment for over 5 days.

- Created an environment using Pandas to simulate the trading environment for the bot.

- Developed several TensorFlow models to learn how to trade following several deep-Q-learning algorithms.

- Created a function to automatically get and normalize data with z-score using Pandas and Gemini API.

AMP Purdue PyTorch, NumPy, Pandas

- Created a TensorFlow model to control a self-driving car without help from LiDAR sensor. The model was based on NVIDIA’s PilotNet paper.

- Trained a PyTorch model to control the car.

- Developed Unity’s simulator to automatically get and normalized images, waypoints, and steering angle data using Pandas and NumPy.

HealthEye Microsoft Azure, Google Map API, React, Flask, MongoDB,

- Developed an IoT system to monitor the number of people and social distancing violations through Raspberry Pie and displayed the number of social distancing violations on a website. Users can pick different locations on Purdue University’s campus to see whether it is safe to go there.

- Developed a full-stack web application to display information for the user using Flask and React.

- Created a MongoDB database instance to store information on Azure from Raspberry Pi and Purdue’s COVID daily information.

Spotify Recommendation System Pandas, NumPy, Sci-kit learn, and Google Colab

- Created a dataset using TF-ids algorithm with Sci-kit learn.

- Implemented K-means clustering to generate top recommendations for users.

- Developed a notebook available for the users.

Skills

Technology: Java, Python, JavaScript, TensorFlow, Keras, PyTorch, NumPy, Pandas, HTML / CSS, MongoDB

Relevant Courses: Problem Solving and Object-Oriented Programming

Contact this candidate