University of Texas at Austin – Post Graduate Program in Artificial Intelligence & Machine Learning
This repository is a consolidated portfolio of 11 end-to-end Data Science, Machine Learning, MLOps, Computer Vision, and GenAI projects completed as part of the Post Graduate Program in Artificial Intelligence & Machine Learning offered by the University of Texas at Austin.
The projects are designed around real-world business problems and span the complete analytics and ML lifecycle — from business understanding and exploratory data analysis to modeling, evaluation, deployment concepts, and actionable insights.
Each project is maintained as a separate folder within this repository for clarity and modularity.
git clone https://github.com/ananttripathi/AI-ML-Projects-UT-Austin.git
cd AI-ML-Projects-UT-Austin
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
pip install -r requirements.txtOpen any project's notebook (e.g. food-delivery-demand-analysis/FoodHub_Project.ipynb) in Jupyter or Google Colab. Each project folder has a readme.md with run instructions and required data files.
AI-ML-Projects-UT-Austin/
├── food-delivery-demand-analysis/
├── personal-loan-propensity-modeling/
├── ab-testing-landing-page-conversion/
├── agentic-news-retrieval-system/
├── retail-sales-forecasting-system/
├── vehicle-resale-business-analysis-sql/
├── visa-approval-prediction-system/
├── wind-turbine-predictive-maintenance/
├── safety-helmet-detection-system/
├── customer-purchase-prediction-mlops-pipeline/
└── medical-knowledge-rag-system/
Each project folder typically contains:
- Business context & problem definition
- Analysis / modeling notebooks
- Key findings and business insights
- A
readme.mdwith how to run, files overview, and key libraries
| # | Project | Folder | Domain |
|---|---|---|---|
| 1 | Food Delivery Demand Analysis | food-delivery-demand-analysis |
Business Analytics |
| 2 | Personal Loan Propensity Modeling | personal-loan-propensity-modeling |
Banking / Marketing |
| 3 | A/B Testing – Landing Page Conversion | ab-testing-landing-page-conversion |
Experimentation / Stats |
| 4 | Agentic News Retrieval System (NewsFindr) | agentic-news-retrieval-system |
GenAI / NLP |
| 5 | Retail Sales Forecasting System | retail-sales-forecasting-system |
Time Series |
| 6 | Vehicle Resale Business Analysis (SQL) | vehicle-resale-business-analysis-sql |
SQL / BI |
| 7 | Visa Approval Prediction System | visa-approval-prediction-system |
Applied ML |
| 8 | Wind Turbine Predictive Maintenance | wind-turbine-predictive-maintenance |
Industrial ML |
| 9 | Safety Helmet Detection System | safety-helmet-detection-system |
Computer Vision |
| 10 | Customer Purchase Prediction – MLOps Pipeline | customer-purchase-prediction-mlops-pipeline |
MLOps / CI-CD |
| 11 | Medical Knowledge RAG System | medical-knowledge-rag-system |
GenAI / Healthcare |
Domain: Business Analytics / Product Analytics
Objective:
Analyze customer ordering behavior, restaurant demand, ratings, and operational metrics to help a food aggregator improve customer experience and operational efficiency.
Key Focus Areas:
- Restaurant demand patterns
- Cuisine preferences
- Order timing (weekday vs weekend)
- Ratings and delivery performance
📁 Folder: food-delivery-demand-analysis
Domain: Banking / Marketing Analytics
Objective:
Predict which liability customers are most likely to purchase personal loans and identify key attributes driving conversion, enabling targeted marketing campaigns.
Key Focus Areas:
- Customer segmentation
- Feature importance & explainability
- Conversion-focused classification
📁 Folder: personal-loan-propensity-modeling
Domain: Experimentation / Statistical Inference
Objective:
Evaluate whether a redesigned landing page improves user engagement and subscription conversion using hypothesis testing at a 5% significance level.
Key Focus Areas:
- Controlled experimentation
- Conversion rate analysis
- Time-on-page comparison
- Language-based behavior analysis
📁 Folder: ab-testing-landing-page-conversion
Domain: GenAI / Natural Language Processing / Agentic AI
Objective:
Build an agentic news retrieval system that uses LLM-based agents to search, retrieve, and summarize news content—demonstrating orchestration of tools, retrieval, and natural language generation.
Key Focus Areas:
- Agentic workflows & tool use
- News search and retrieval
- Query understanding and response generation
📁 Folder: agentic-news-retrieval-system
Domain: Time Series Forecasting
Objective:
Forecast quarterly sales revenue across retail outlets to support inventory planning and regional sales strategies.
Key Focus Areas:
- Sales trend analysis
- Forecasting techniques
- Business planning implications
📁 Folder: retail-sales-forecasting-system
Domain: SQL / Business Intelligence
Objective:
Generate a quarterly executive business report analyzing sales performance, customer feedback, and operational KPIs using SQL.
Key Focus Areas:
- Multi-table SQL queries
- KPI computation
- Executive-level reporting
📁 Folder: vehicle-resale-business-analysis-sql
Domain: Applied ML / Decision Science
Objective:
Predict U.S. visa certification outcomes and identify significant drivers influencing approval or denial decisions.
Key Focus Areas:
- Policy-oriented classification
- Feature impact analysis
- Decision-support insights
📁 Folder: visa-approval-prediction-system
Domain: Industrial ML / Cost-Sensitive Learning
Objective:
Predict wind turbine generator failures using sensor data to optimize inspection, repair, and replacement costs.
Key Focus Areas:
- Cost-sensitive classification
- False positive vs false negative trade-offs
- Maintenance strategy optimization
📁 Folder: wind-turbine-predictive-maintenance
Domain: Computer Vision
Objective:
Develop an image classification model to automatically detect helmet compliance in hazardous work environments.
Key Focus Areas:
- Image preprocessing
- CNN-based classification
- Deployment readiness
📁 Folder: safety-helmet-detection-system
Domain: MLOps / CI-CD / Deployment
Objective:
Design and deploy an end-to-end MLOps pipeline to automate data processing, model training, evaluation, and deployment for customer purchase prediction.
Key Focus Areas:
- Modular ML pipelines
- CI/CD using GitHub Actions
- Model deployment & monitoring concepts
📁 Folder: customer-purchase-prediction-mlops-pipeline
Domain: GenAI / Healthcare AI
Objective:
Build a Retrieval-Augmented Generation (RAG) system using medical manuals to support accurate, source-grounded clinical decision-making.
Key Focus Areas:
- PDF ingestion & chunking
- Vector search & retrieval
- Hallucination reduction
- Trustworthy AI responses
📁 Folder: medical-knowledge-rag-system
- Programming: Python, SQL
- Data Analysis: Pandas, NumPy
- ML / DL: Scikit-learn, TensorFlow, Keras
- Visualization: Matplotlib, Seaborn
- GenAI: RAG pipelines, embeddings, vector stores
- MLOps: GitHub Actions, CI/CD pipelines
- Platforms: Jupyter Notebook, Google Colab
- Version Control: Git, GitHub
- Demonstrate end-to-end data science problem solving
- Showcase business-first ML & AI thinking
- Highlight deployment-aware and production-ready approaches
- Serve as a comprehensive discussion base for interviews
Suggested GitHub topics: machine-learning data-science mlops computer-vision rag ut-austin portfolio
These projects were developed for educational and demonstration purposes as part of an academic program. All datasets used are either publicly available or provided strictly for learning and non-commercial use.
⭐ Please explore individual project folders for detailed implementations, analyses, and insights.