Computer Science, University of Minnesota - Twin Cities (Graduation: May 2026)
Undergrad Researcher @Minnesota NLP
Currently doing undergraduate research on Reasoning Trajectories, Knowledge Discovery, and Test-Time Scaling in Language Models at Minnesota NLP under the supervision of Professor Dongyeop Kang.
Open for any work opportunities for Summer 2026 and beyond.
I have research and industry experience in Machine Learning and Data Science. And a strong foundation in Computer Science and Software Engineering principles and understand and practice software development life cycle and architectural design.
Designed a custom architecture with separated Time and Channel attention blocks, connected via bi-directional cross-attention. Conducted comparative studies on augmentation strategies and adopted a self-supervised contrastive objective, trained on 2x Nvidia DGX Spark (GB10).
View project →Implemented Word2Vec (Skip-Gram, Negative Sampling), Neural Probabilistic Language Models, and N-Grams to build foundational language modeling concepts from the ground up.
View project →Developed an autonomous RAG-powered Agentic system that generates marketing campaign architectures by integrating past campaign data, client-specific market analysis, and research insights.
Built an LLM-powered resume extraction pipeline using a multi-threaded producer-consumer architecture for efficient processing, aggregating multi-source candidate data into Google Sheets with structured insights.
Led a customer retention project, delivering a proof of concept with clustering (K-Means, GMMs) and churn prediction models (XGBoost, LSTMs) for customer segmentation and retention insights.
Developed a PyTorch-based image classifier based on MobileNetV3 with 96% accuracy on 12,000 images and created an Android app for content blocking.
Trained an LSTM Network on 1 million texts from Twitter interactions and movie reviews with 87% accuracy, and deployed it as an API endpoint.
View project →