This repository contains portfolio projects for the AI Engineer Career Path, demonstrating practical applications of neural network engineering, LLM operations, AI applications, and AI agent development.
Classifying Banking Intent From Customer Queries
This project demonstrates an end-to-end AI engineering workflow for an NLP classification task, comparing traditional neural networks with modern transformer-based models finetuned with LoRA (Low-Rank Adaptation).
Project Overview: Build a production-ready AI system that classifies customer banking intent from queries across 77 intent classes. This project showcases:
- Modern AI Architectures: Compare Multi-Layer Perceptron (MLP) classifiers with LoRA-finetuned RoBERTa transformers
- Production-Ready Systems: Implement comprehensive PII (Personally Identifiable Information) protection for sensitive financial data
- End-to-End Pipeline: From exploratory data analysis through model training, evaluation, and responsible AI implementation
What You'll Build:
-
Exploratory Data Analysis
- Explore and visualize customer query texts
- Analyze distribution patterns across intent classes
-
Model Development
- Build and train MLP classifier with text preprocessing pipelines
- Finetune RoBERTa transformer using parameter-efficient LoRA
- Compare traditional vs modern architecture performance
-
Model Evaluation
- Establish baseline metrics across 77 intent classes
- Analyze per-class performance improvements
- Compare training efficiency and computational requirements
Prerequisites:
- Python fundamentals
- Data Science: Pandas, NumPy, Matplotlib, Seaborn, scikit-learn
- Machine Learning: Classification metrics, train-test splits, evaluation
- AI/Deep Learning: Neural networks, PyTorch, Transformers (Hugging Face), LoRA
- Text preprocessing: Data cleaning, regex patterns, PII detection, hashing
View the Engineer Neural Networks project →
This career path uses Jupyter Notebook and Git version control. If you need help with setup:
- Command Line Interface Setup
- Introducing Jupyter Notebook
- Setting up Jupyter Notebook
- Getting Started with Jupyter
- Getting More out of Jupyter Notebook
Need a refresher on Git? Check out:
Feeling stuck? Try these strategies:
- Google your question: Check StackOverflow and Dev.to for community solutions
- Read the documentation: Carefully review documentation for languages and libraries you're using
- Rubber ducking: Explain your problem to a friend or colleague—you'll often figure it out as you talk through it
This is a learning repository for the AI Engineer Career Path. Each project is designed to be completed independently as part of your portfolio development.
These projects are for learning/demo purposes.
Ready to build production-ready AI systems? Start with the Engineer Neural Networks project and demonstrate your end-to-end AI engineering skills!