QLoRA LLM Fine-Tuning Pipeline

• Focus: Applied AI, model fine-tuning, experiment tracking
• Tech Stack: Python, Hugging Face Transformers, PEFT, TRL, BitsAndBytes, DVC, MLflow

Designed and implemented a fine-tuning pipeline for a 7B parameter language model using QLoRA-style parameter-efficient training. The workflow includes custom data extraction, model versioning, and experiment tracking.

This project highlights practical AI engineering: reproducible experiments, controlled model iterations, and infrastructure that supports secure and scalable model development.