Glen Miasnychenko

Jr. AI EngineerOpticsPlanet - Northbrook, IL - 2025-to date

Transformer-based Recommendation System Development

• Researched transformer-based recommendation models including SASRec and BERT4Rec architectures
• Developed prototype systems and trained them on MovieLens dataset to compare performance metrics
• Implemented production-ready BERT4Rec training pipeline deployed on cloud infrastructure

Technologies used: BERT4Rec, SASRec, PyTorch, Google Cloud, AWS, Transformers

AI Engineer Intern OpticsPlanet - Northbrook, IL - Summer 2024

Retrieval-Augmented Generation (RAG) System Development

• Integrated vector storage with pre-trained sentence-transformer models for document chunking, embedding, and efficient search
• Customized document processing pipeline for Markdown-based Confluence articles
• Built a conversational module utilizing various LLM
• Refined indexing and embedding logic to optimize performance, enabling real-time query responses

Technologies used: PyTorch, OpenSearch, Docker, Confluence API, Flask, Llama3 models, OpenAI API, Sentence-transformers

Roosevelt University, College of Arts
and SciencesChicago, IL, December 2024

Bachelor of Arts in Actuarial Science with minor in Finance

Relevant Coursework: Regression & Time Series, Probability Theory, Linear Algebra, Mathematical Statistics, Financial Mathematics, Data Mining

NN for Microplant Classification Roosevelt University - Spring 2024

• Collaborated with the Field Museum to classify microplants using a dataset of images provided by the museum
• Developed and implemented a convolutional neural network (CNN) from scratch to perform the classification
• Handled all aspects of data processing, including parsing, cleaning, and downloading images using Python scripts
• Achieved peak validation accuracy of 65.84% using advanced techniques such as data augmentation, dropout layers, and regularization
• Automated hyperparameter tuning with Hyperband to optimize model performance

Bookkeeping Application Personal Project - Fall 2023

• Developed a full-stack personal bookkeeping application using Django framework
• Implemented key features such as transaction tracking, account balance updates and financial summaries with user-friendly interfaces
• Migrated the application to run on a Kubernetes (k3s) cluster, improving scalability and enabling smoother deployment
• Utilized MySQL as the database backend, configured distributed environments, ensuring data consistency and optimized performance
• Dockerized the application for containerized deployment and streamlined application updates

Walmart Sales Analysis Roosevelt University - Fall 2023

• Lead a project on modeling the Walmart quarterly sales
• Employed various statistical methods such as Decision Trees, Best Subset Selection (Ridge and Lasso), Forest Model for comprehensive data analysis
• Determined the relationship between the key macroeconomic factors and the Walmart sales

Research on Robinta pseudocaccia Prague University - Spring 2023

• Assisted with statistical analysis of a peer-reviewed research
• Conducted the Tukey statistical test with appropriate checks
• Performed the analysis using SPSS Statistics and later recreated using R

• Programming Languages: Python, R, Golang
• Frameworks: Django, Flask
• Data & ML: Deep Learning, Data Preprocessing, PyTorch, Keras
• DevOps: Kubernetes, Docker, CI/CD, Linux Server Administration
• Databases: MySQL, OpenSearch, Elasticsearch
• Cloud: GC, AWS, Oracle Cloud
• Tools: Git, Excel, SPSS Statistics

• Github
• Linkedin