Support Ticket Triage Platform
A support-ticket triage platform with a Gradio UI and MLflow experiment tracking for model iteration.
Problem
Support teams waste time manually routing and prioritizing incoming tickets before an agent can even start solving the problem.
Solution
A triage classifier with a Gradio UI for real-time predictions and MLflow tracking, so model versions can be iterated on with confidence.
Overview
Every minute a support ticket sits unrouted is a minute an agent isn’t solving it. Manual triage is one of the most avoidable delays in a support workflow.
How it works
A triage classifier predicts ticket category and priority the moment a ticket arrives, exposed through a Gradio UI for real-time testing and demoing. Model iterations are tracked in MLflow, so a new model version can be compared against the previous one before it’s trusted with real tickets.
Result
Tickets get routed automatically instead of sitting in a queue waiting for a human to read and categorize them first.
Further reading
See From Notebook to Data Product for more on why version tracking and a usable UI matter here.
More projects
Product Recommendation Engine
An e-commerce recommendation engine combining a from-scratch ALS implementation (NumPy/SciPy) with content-based scikit-learn filtering.
Marketing Analytics Data Pipeline
An end-to-end marketing analytics data pipeline that ingests, transforms and models campaign data for reporting.
Retail Analytics Warehouse
A dbt + DuckDB ELT pipeline over synthetic retail data, with 35 automated data-quality tests and a Streamlit dashboard.