Ken Vellian

Data Science Portfolio - Ken Vellian

Projects

Retrieval-Augmented Generation System for the Harry Potter Corpus - A Complete End-to-End Private RAG Pipeline (November 2026)

Visit Project
Topic: Decision Sciences (Python, Data Scraping, Cleaning, Preprocessing, RAG, LLMs, Embedding models)
This project develops a complete end-to-end Retrieval-Augmented Generation (RAG) system using the Harry Potter book corpus as a stand-in for enterprise documents. The goal is to simulate how companies deploy Large Language Models (LLMs) over their private, internal data while maintaining privacy, traceability, and retrieval accuracy. The system includes every major pipeline component: web scraping, text cleaning, chunking, embedding generation, hybrid retrieval with blending weights, reranking, evaluation using custom question-answer datasets, grid search over hyperparameters, and a fully deployable Streamlit web application.

Bizi Mart Tobacco Reorder Point List (September 2024)

Visit Project
Topic: Decision Sciences (Python, Excel)
This project aims to create a color-coded tobacco product shopping list for Bizi Mart that highlights levels of urgency to reorder (Red, Yellow, Orange, Green).

Runtime Comparison of SQL Queries vs. Functions on Twitter Databases (March 2024)

Visit Project
Topic: Database Processing for Large-Scale Analytics (Python, SQL, JSON)
This project aims to populate databases and retrieve information in various methods using this Twitter data and compare their performances.

Visualizing Texas Real Estate Trends 2024: 500 Listings 🏠 (March 2024)

Visit Project
Topic: Data Visualization (R)
Collaborative Project
This project explores what the average single-family home in Texas could look like.

Predicting Penguin Gender Through Their Characteristics (November 2023)

Visit Project
Topic: Fundamentals of Data Science (R)
This project explores the accuracy of different machine-learning algorithms to predict penguin gender.

M.S., Data Science | DePaul University | (Expected December 2025)

B.S., Info. Systems & Decision Sciences | Louisiana State University

Operations Manager @ Bizi Mart | Antioch, CA (October 2022 - August 2023)

Implemented the LMS-POS system to scale operations and process a higher volume of transactions, which boosted sales by 43.78% in 1 month.
Improved the report system by tracking products sold by vendors to make more informed decisions when ordering.
Utilized SEO to market the business through Google Maps and Yelp, and managed daily accounting through Excel.
Managed vendor relationships, negotiated contracts to reduce costs, and coordinated with employees to maintain product availability and customer retention.

Client Services Consultant @ Oracle Textura | Pleasanton, CA (Remote) (April 2021 – October 2022)

Drive enterprise-wide SaaS adoption of Textura through pre-sales consulting, client implementations, and ongoing post-sales support, while managing 9 key accounts.
Collaborated with C-level executives, project managers, and accountants to gather business requirements, streamline manual workflows, deliver custom software demos, and execute implementation plans.
Documented user stories and collaborated with the development team to test the quality of new product features, then deployed rollout plans for clients to use upon release.
Resolved a wide range of client support tickets, addressing issues such as lien waiver management, data discrepancies, payment disbursements, troubleshooting integration issues (CMiC, Sage 300, Viewpoint), and onboarding subcontractors. Collaborated with cross-functional teams to solve complex problems and deliver timely solutions.