I build intelligent systems that turn data into action—from LLM platforms used by 500+ employees to ML solutions driving millions in savings.
About
I'm a data scientist and machine learning engineer focused on building practical systems that solve real problems. At Dot Foods, I've worked on everything from a company-wide LLM platform used by 500+ employees to optimization tools projected to save millions annually. Before that, I did research at Southern Illinois University Edwardsville in deep learning and genomic data analysis. I'm currently pursuing a Master of Computer Science in Data Science at the University of Illinois Urbana-Champaign while working full-time, and I'm motivated by curiosity, continuous learning, and building tools that make a real impact.
Experience
From production AI systems at enterprise scale to academic research in deep learning and genomics.
Projects
Production systems, ML models, and research that delivered real-world impact.
Architected a full-stack Azure AI platform backed by Postgres and Redis, serving as the unified interface for all internal agentic use cases. Designed to scale across the entire organization with governed access, extensible tool integrations, and production-grade reliability.
Built an end-to-end fuel optimization system combining Resource-Constrained Shortest Path algorithms with Generative AI. Deployed as an Azure Function App with a Snowflake data pipeline and APIM API layer, optimizing fuel stop decisions across the distribution network.
Developed and deployed a LightGBM classification model to predict early-stage attrition in entry-level warehouse roles. Achieved 83% recall with an 8-week lead time on terminations, enabling proactive retention interventions before they become costly.
Built an AI-powered agent using GPT with RAG and OCR capabilities, connected to both DB2 and Snowflake data sources. Provides real-time support for customer service representatives, surfacing relevant order history, documentation, and policy context instantly.
Designed a personalized job matching system powered by GPT text-embedding-3-large for semantic similarity scoring and GPT-4o-mini for conversational recommendations. Matches employees to internal roles based on skills, experience, and career aspirations.
A daily music puzzle game blending Wordle-style mechanics with music knowledge. Features two game modes—Daily Cue (guess the song from a scenario prompt) and Setlist Crisis (pick the best song for chaotic prompts). Puzzles are auto-generated daily via OpenAI with enriched song metadata.
Comparative analysis of deep convolutional neural networks and traditional ML methods for disease segmentation and classification in tomato and rice leaves. Evaluated AlexNet, ResNet50, InceptionResNetV2 against Random Forest and K-Means across illumination conditions.
Skills
The languages, frameworks, and platforms I use to build and ship.
Education & Research
Evaluated deep learning (AlexNet, ResNet50, InceptionResNetV2) and traditional ML (Random Forest, K-Means) methods for crop disease classification. Found DCNNs generally outperformed unsupervised ML methods, with dataset-specific performance variations under different illumination conditions.
Conducted research in the NIH All of Us Research Workbench to identify single-nucleotide polymorphisms (SNPs) associated with Autism Spectrum Disorder comorbidities, working with large-scale genomic datasets using Hail for variant analysis.
Community
Teaching, mentoring, and building communities around technology.
Contact
Always open to discussing data science, AI engineering, research, or new opportunities.