Spaces:
Runtime error
Runtime error
File size: 1,528 Bytes
d3c90c1 388ac2e d3c90c1 388ac2e d3c90c1 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 | ---
title: Blocks AI Assistant
emoji: 🏢
colorFrom: red
colorTo: red
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
pinned: false
---
# Blocks Assistant – Hybrid AI + Analytics
This Space hosts a fully self-contained AI Operations Assistant built for backlog mitigation,
block analysis, and fiscal-quarter trend detection.
### Key Capabilities
- 📊 Block summary (Pricing, Credit, Delivery, Ship Debit)
- 🧮 Customer exposure analytics
- 🔁 Dynamic FAQ learning based on user queries
- 🧱 Multiple block order detection
- 📅 Fiscal-year quarter-wise trend insights (Apr–Mar FY)
- 🧠 Phi-2 LLM rewriting layer for polished operational insights
- 🔒 Zero external APIs – all analytics & AI happen inside the Space
### How it works
1. Upload your backlog file:
**Backlog_updated.xlsx** (required)
2. The assistant computes analytics:
- Block exposure
- Order-level details
- Highest exposure customers
- Multi-block orders
- FY quarter trends
3. The LLM refines the insights:
- Business-professional tone
- Zero hallucinations (numbers are not changed)
- Concise operations-style recommendations
### Technology Stack
- Flask backend
- Pandas for analytics
- Transformers + Torch (Phi-2 model)
- HTML + JS + CSS frontend (custom UI)
- Runs fully on CPU
### No external API calls
- No OpenAI
- No remote LLMs
- No data leaves Hugging Face
Designed for operational reliability.
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
|