Production-grade machine learning pipeline for customer churn prediction with modular components for data preprocessing, model training, and streaming inference. This third repo in a series builds on previous work to create a deployment-ready prediction system with comprehensive configuration, evaluation metrics, and scalable architecture.

A local‑first AI pipeline that generates coherent images, videos, stories, dialogues, and spatial audio from an augmented storyboard. It structures prompts, ensures narrative continuity, automates sound spatialization, and provides a reproducible workflow for fast, high‑quality multimedia creation.

Modular pipeline toolkit for DCCs

MLOps producer hub for medicine pack identification. Built with Streamlit, YOLOv8-segmentation, and DINOv2. Featuring automated vector database synchronization with AWS S3 and activity logging system

A gated, artifact-based operating system for AI film and narrative production with continuity control, cost governance, and model-agnostic routing.

End-to-end Azure Databricks retail data engineering project using Medallion Architecture (Bronze, Silver, Gold). Implements Auto Loader, Unity Catalog, Delta Lake, SCD Type 1 & 2 dimensions, and Fact Orders for analytics-ready star schema modeling.

Modern batch data pipeline using AWS Glue, S3, Snowflake, and dbt. Showcases secure ingestion, external stages, layered transformations, dbt testing, and production deployment with environment separation.