Melanie Ato Yadgar

I am a computer engineer that has experience in designing machine learning algorithms in an academic setting and for personal enjoyment. I have also designed neural nets models for school projects. My plan is to further my education at UW-Madison since I have been accepted into the Masters' program for Machine Learning and Signal Processing.

Short-term Goal: Gather industry experience in machine learning this Summer to take into masters’ degree machine learning and signal processing program this Fall.

Long-term Goal: Utilize affluence with PyTorch and neural net modeling in data analytics industry.

Proficient in: Java, JavaScript, Python, Matlab, SystemVerilog, C

Fundamental Knowledge in: HTML, CSS, Verilog, Perl, VHDL, C++, Bash, MySQL

• Filesystem Online (Spring 2024)

Developed a full stack product which was an in-browser, online file management system, similar to Google Drive.

Tech stack consisted of a docker image running the MySQL database on a Linux server which connected with a Java-written API with calls associated in a JS frontend.

User accounts were authenticated by Google.

• Football Vegas Betting Predictor (Fall 2023)

Created a classifier that optimized over-under winning score predictions prevalent in Las Vegas sports betting; used a 5-fold crossvalidation training set with parameter dictionaries defined for Support Vector Machine (SVM), KNN classifier, Logistic Regression model, and Linear regression model.

Ensemble models were also trained to further optimize the result.

Highest accuracy achieved was greater than 61%.

• Weather Predicting Convolutional Neural Network (CNN) (Spring 2023)

Worked in a team of four people to create a convolutional neural network, designing it layer by layer using empirically proven practice with defining hyper-parameters. Ended up with a three-layer CNN that could predict weather conditions from images with over 87% accuracy. Used a PyTorch stack.

• WICFA22 (Fall 2022)

Designed a Single Cycle Processor from a Microprocessor Systems Design Course, all made with a SystemVerilog and Bash code base.

Project implemented the PC, Memory, ALU, and register file of a single cycle CPU using course-designed ISA.

• Electric Bike (Spring 2022)

Motor for an electric bike designed using FPGA technology paired with a SystemVerilog code base.

I organized the design project by leading a four-person team.

Circuit design was validated and tested using the crankset of a bicycle.

• Memory Manager (Spring 2021)

I programmed the course files for a custom memory manager in C for allocating and freeing memory.

All executables were written in C.

This developed intimate understanding of native memory allocator of ANSI standard C.

• Zork (June 2018)

Programmed a Command Line Interface (CLI) Choose-Your-Own-Adventure Game

Java comprised the main executable.

I created a room-based dungeon crawler with mini-bosses and a final boss.

• Online Diary (June 2017)

I designed an in-browser, online diary where one can write, save, edit, and delete text entries.

JavaScript was the main executable.

Diary was hosted by a third-party website.

• Jedi Battle (June 2017)

I developed this game for the Xbox 360 Kinect platform.

Scratch was main executable.

I led and coordinated this project among a 3-person team, acting as the main developer.

• Utilized Statistics and Programming based knowledge to write a program that analyzed the balance of custom-made classes for Dungeons & Dragons (DnD).

• Aero-Flex Pre-Apprenticeship Certificate of Completion: Training from partner with Northrop Grumman and various other aerospace companies to meet South Bay Workforce Investment Board approved standards for corporate cooperation.
• National Academy of Future Scientists and Technologists Award of Excellence - Awarded in 2018 to the selection of John C. Mather, Ph.D., executed in Boston, Mass.
• Received several certificates of academic excellence from high school, including in Biliteracy.