Home
Resume
Education
Skills
Projects
Research
Experience
Shop
Contact

Genova Mongalo

My name is Genova Mongalo, I am a summa cum laude graduate with a Bachelor of Science in Computer Science from the University of Missouri-Kansas City (UMKC). I am currently pursuing a Master's degree in Computer Science with an emphasis on Artificial Intelligence at Georgia Tech, beginning in August 2025. My primary focus is on AI/ML Engineering and Cybersecurity applications.

During the summer of 2025, I worked as a Federal AI/ML Engineer Contractor with the AFRL Sensors Directorate Internship Program at the University of Dayton. This role involves pioneering the development of generative AI models that enable CNN classifiers to recognize real objects when trained on generated data, utilizing transfer learning frameworks to address AFRL/DoD priorities.

I am passionate about applying machine learning and artificial intelligence to solve real-world problems, particularly in cybersecurity and defense applications. I believe my strong foundation in AI/ML, coupled with my hands-on experience in cutting-edge research, positions me to contribute effectively to innovative projects in the tech industry.

Resume

Educational Background

University of Missouri-Kansas City (UMKC) — Kansas City, MO

Bachelor of Science in Computer Science, summa cum laude
August 2021 – December 2024

Georgia Tech — Atlanta, GA

Master's in Computer Science - Emphasis on Artificial Intelligence
Starting August 2025

Graduate Coursework

Georgia Tech — M.S. in Computer Science (AI emphasis)

Deep Learning
- Neural network architectures: CNNs, RNNs, and transformers
- Training, optimization, and regularization of deep models at scale
Reinforcement Learning
- Markov decision processes, value-based and policy-gradient methods
- Reward design and sequential decision-making for autonomous agents
Machine Learning
- Supervised and unsupervised learning algorithms and applications
- Model training, evaluation, and generalization
Artificial Intelligence
- Search, knowledge representation, planning, and reasoning
- Foundational AI techniques underpinning modern ML systems
Brain & Cognitive Science
- Computational models of perception, memory, and learning
- Neuroscience principles that inspire AI architectures

Core Foundations

Algorithms & Complexity
- Algorithm design and analysis of computational complexity
Data Structures
- Efficient organization, storage, and access of data

Skills

Programming Languages

Python, C / C++, Java / C#, SQL (Postgres), JavaScript, HTML/CSS, LISP

Frameworks & Technologies

Machine Learning
React
Node.js
WireShark
FastAPI

Developer Tools

Git, VS Code, Visual Studio, PyCharm, Eclipse, Anaconda, Unity, Google Cloud Platform

Libraries

pandas, NumPy, Matplotlib, PyTorch, tkinter

Specialized Skills

Generative AI & GANs
Large Language Models
Transfer Learning
CNN Classifiers
Cybersecurity Applications
Parallel Computing (Slurm)

Spoken Languages

English, Spanish, and Amharic

Projects

Federated Blockchain — Model Registry (2025–2026)

A federated-learning model registry on Ethereum that lets nodes share and audit machine-learning models without putting large weights on-chain. Solidity smart contracts (FederatedModelStorage, FederatedLedger, and an ERC-20 FederatedToken), deployed with Hardhat and tested on Ganache, record each model's IPFS content ID on-chain. Trained models — including the ALBERT ransomware classifier from my research — are exported as self-describing bundles, uploaded to IPFS, and registered via either a Node.js script or a React/Vite frontend with MetaMask wallet integration. Includes an automated Hardhat test suite verifying a model CID round-trips through all three contracts.

Racket Parser (2024)

A custom parser built in Racket for a small artificial programming language inspired by FORTRAN. It reads .txt source files, validates syntax based on a defined grammar, and reports precise error lines when invalid constructs are encountered. This project demonstrates understanding of language grammars, recursive parsing techniques, and working with DrRacket tools to implement language analysis.

Phonebook Management System (2024)

A Java-based phonebook management application that integrates with Dolphin SQL for persistent storage. It supports user registration with secure password handling, contact addition/update/deletion, search functionality, and CSV export. Developed using Eclipse, this project highlights practical database integration, user interaction design, and CRUD operation management in a desktop application.

SOS Game Implementation (2023)

An implementation of the classic SOS paper-and-pencil game designed to explore software development workflows using Scrum methodology. Written in Python, this project reflects iterative development practices, automated testing, and simple game logic implementation. It’s a great example of applying agile principles to deliver a functional interactive game while learning team processes.

Research

Ransomware Detection for Critical Infrastructures (2024–2026)

A comparative study of deep-learning and transformer models for detecting and classifying ransomware from Windows API-call features, aimed at Industrial Control Systems. Each model predicts at three levels simultaneously — binary (benign vs. malicious), coarse family group, and specific family (12 classes) — and is evaluated under three feature-engineering setups (baseline, autoencoder, and K-Means clustering). The pipeline compares a DNN, a 1-D CNN, and RoBERTa- and ALBERT-style transformers against classical baselines, using chi-squared feature selection to reduce ~31,000 features to the most informative 1,000 and class-weighting to handle rare families. The reproducible PyTorch/scikit-learn pipeline reaches ~98% accuracy on binary detection, with early stopping and weight sharing (ALBERT) used to control overfitting on the small dataset. GPU-accelerated on Apple Silicon via Metal.

Work Experience

AFRL Sensors Directorate Internship Program, University of Dayton — Federal AI/ML Engineer Contractor

May 2025 - August 2025

Pioneering the development of generative AI models enabling a CNN classifier to successfully recognize real objects when trained on generated data when utilized in transfer learning framework, addressing AFRL/DoD priorities. Using PyTorch, adapting state of the art classifiers to handle novel imaging modalities. Utilizing slurm to enable large batches of parallelized computation. Creating an innovative game theory approach to GAN training, significantly improving the results of the generator's development. The generative model's output is expected to lead to an increase of 50 percentage points in the downstream classifier compared to the baseline.

NSF REU AI-Empowered Cybersecurity, University of Missouri Kansas City — AI/ML Engineer Intern

June 2024 - December 2024

Engineered a robust Large Language Model to detect ransomware threats within Industrial Control Systems. Achieved 99% accuracy for binary classification and 85% accuracy for family classification scale respectively. Research was accepted and presented at IEEE Big Data 2024 in Washington, D.C. as the solo author with mentor guidance.

NSF REU in Consumer Networking, University of Missouri Columbia — Research Intern

May 2023 - July 2023

Designed a model that detects/predicts cybersickness using a transformer machine learning model. Succeeded with an accuracy of 85% from a 1-10 cybersickness severity scale.

NASA Missouri Space Grant Consortium — Research Intern

December 2021 - April 2022

Engineered an AI-driven simulation of space-bound transmissions as part of a NASA-funded initiative analyzing signal propagation and identifying optimal frequency bands under varying conditions. Developed an interactive interface to visualize transmission strength and provide adaptive frequency recommendations based on environmental parameters.

SPCN - 2020 and IEEE Brain Initiative BDBC Conference, Taiwan — Web Master

August 2020 - September 2020

Set up conference website for the 5th International Symposium on Audio and Video Signal Processing in the Context of Neurotechnology, with reliable communication and updates to the website’s needs. The conference was conducted remotely through multiple worldwide locations.