Josh Janzen

About Me

As a Data Science Director and member of the GenAI Center of Excellence, I specialize in transforming business vision into measurable outcomes through AI and analytics initiatives. My focus is on identifying and implementing high-impact AI solutions that drive significant business value and accelerate digital transformation.

Featured Projects

AI Roster Optimization

Built a comparative study between OpenAI's o1 model and Google OR-Tools for optimizing basketball team rotations, demonstrating different approaches to complex scheduling constraints.

Key Features:
- Interactive web interface using Streamlit
- Side-by-side comparison of LLM vs OR-Tools solutions
- Complex constraint handling
- Visual rotation schedule representation
Architecture: FastAPI backend, Streamlit frontend, Docker containerization

Technologies: Python, FastAPI, Streamlit, OpenAI API, Google OR-Tools, Docker

AI-Powered Math Learning Assistant

Developed an intelligent math tutoring system using LangGraph for reliable and controlled question generation workflows.

Key Features:
- Grade-adaptive questioning system (Pre-K to 12th)
- Sophisticated state machine for question generation
- Comprehensive validation workflow
- Real-time hint generation
Architecture: LangGraph state management, Docker containerization

Technologies: LangGraph, Streamlit, Docker, StateGraph

RAG-Powered Document Intelligence System

Built a production-grade RAG (Retrieval Augmented Generation) application with separated frontend/backend architecture, enabling intelligent document conversations with custom PDFs.

Architecture: FastAPI Backend, Streamlit Frontend
Key Features:
- Hugging Face embeddings with FAISS vector storage
- Docker containerization with Make automation
- Production-ready architecture with clear separation of concerns

Technologies: Python, FastAPI, Streamlit, LangChain, Docker

Impact & Experience

AI & Data Science Leadership

Director of Data Science - C.H. Robinson

GenAI initiatives including LLM implementation for document processing
Developed and monetized data science applications connecting customer shipment data to transportation optimization
External-facing Data Science application evaluating supply chain efficiency

Technical Innovation

Principal & Senior Data Scientist Roles

Architected modern data solutions using Docker, FastAPI, and cloud technologies
Led implementation of recommendation systems and supply chain optimization tools
Pioneered data product development across multiple Fortune 500 companies

Technical Expertise

AI & Machine Learning

Large Language Models (LLMs)
Retrieval Augmented Generation (RAG)
Natural Language Processing

Technologies

Python, Docker, FastAPI
Azure & Google Cloud Platform
LangChain, Transformers

Leadership

AI Strategy & Implementation
Team Development & Mentoring
Product Monetization
Digital Transformation

Conference Presentations

AutoML: Automating Data Science

MinneAnalytics Data Tech 2019

Explored how AutoML tools can enhance data scientist productivity while discussing the balance between automation and expertise. Demonstrated practical applications of AutoML in real-world scenarios.

Key Points:

• AutoML for data preparation and model tuning
• Impact on data science team efficiency
• Live tool demonstration

ETL Clickstream to Data Product APIs

MinneAnalytics MinneFrama 2018

Demonstrated a four-step process to transform clickstream data into production-ready ML models and APIs, requiring no backend engineering support.

Technologies:

• Apache Spark
• Cloud Computing
• REST APIs

Algorithmic Product Recommendations

MinneAnalytics FARCON 2015

Deep dive into the mathematics behind product recommendation systems, including live demonstrations of algorithms in action.

Focus Areas:

• Recommendation algorithms
• Customer experience enhancement
• Live system demonstration

Competitive Spirit

Professional Natural Athlete

NGA Pro Male Physique Competitor since 2015

Key Achievements:

2015 NGA Titan Classic - 1st Place Open Physique (Pro Card)

Data Science Director