8-minute public review

Help validate a data-quality LLM agent.

Try the public demo, inspect the support-ticket failure, and leave one GitHub issue with what was useful, confusing, or missing. Your review helps turn this project into public, resume-safe evidence.

Open Demo Submit Feedback Open Review Request

Current feedback

0 public items

External feedback starts at zero until linked GitHub issues prove otherwise.

Current users

0 confirmed

A confirmed user must state that they tried the demo or ran the repo.

Public footprint

13 issues · 1 fork

GitHub issue count increased after opening the public review request.

CI evidence

147 tests

The repo verifies agent behavior, API routes, safety boundaries, and evidence artifacts.

What to check

The demo analyzes an anonymized support-ticket export. The expected result is a failing quality report with duplicate IDs, missing routing fields, negative amounts, and outlier evidence.

Inspect Report Snapshot Read Case Study Copy Review Invitations Request External Runs

Open the public demo

Skim the report snapshot and confirm whether the data-quality failure is understandable.

Open Demo

Check the AI-agent signal

Look for tool-calling, read-only data boundaries, structured evidence, verifier guardrails, and fallback behavior.

Agent Capability Matrix

Leave one public note

Write what worked, what confused you, and whether the project feels credible for an AI Engineer internship review.

Submit GitHub Feedback

Counts toward external_feedback_items.

Optional: confirm a run

If you ran the local repo or replayed the CSV/PostgreSQL path, submit a redacted replay note.

Submit Replay Evidence

Counts toward reproducible_feedback_items.

Safety boundary

Please do not post raw customer data, secrets, private emails, addresses, tokens, production rows, or private company names. Redacted summaries are enough.

View Evidence Funnel View Public Metrics