.jpeg)
Overview
To explore and prototype innovative agentic AI workflows that demonstrate the power of autonomous AI agents working together to solve complex business challenges across fashion technology and enterprise data management.
Use Cases
SQL Database Chat
Objective
Enable non-technical users to query relational databases using plain English and get human-friendly insights without writing SQL.
Description
This project uses Generative AI to create an intelligent SQL database assistant. Users can ask natural language questions like "What was the highest sale last month?" and receive both the corresponding SQL query and a natural-language summary of the results. It simplifies data access for business users, analysts, and decision-makers.
Key Features
Natural language to SQL query conversion
Result summarization in plain English
Error handling for ambiguous or invalid queries
Tech Stack
Front End
Streamlit
LLM Framework
LangChain
Pydantic
Autogen
LLM Provider
OpenAi
Database
Supabase
Architecture Overview
User inputs question
LangChain processes prompt using LLM
Generates SQL query
Executes on live database
Returns query + natural language result
Benefits
Empowers non-SQL users to interact with data
Saves time for analysts and business users
Reduces reliance on technical teams for routine queries
Bridges the gap between raw data and decision-making
Challenges
Complex query generation (e.g., joins, nested logic)
Schema understanding in large databases
Maintaining SQL safety and injection prevention
Handling vague or overly broad questions
VendorBoard – Smart & Simple Sales Tracker App
Key Features
JWT-based Secure Authentication: Login/signup system to protect user
data.
Sales Entry Panel: Enter product name, quantity, and price with auto-
calculation of totals.
Sales Analytics: Visual insights with graphs and metrics for weekly,
monthly, and yearly sales.
Sales History: Filter and view previous sales with download and edit
functionality.
Excel Report Download: Export filtered sales history as .xlsx files.
Responsive Dashboard: Mobile/tablet-friendly interface with collapsible
sidebar.
Description
VendorBoard is a modern, user-friendly sales tracker that enables users to
record their daily product sales, analyze trends, and download historical data in Excel format. With JWT-based secure login, users access a clean dashboard where they can switch between sales entry, analytics, and history—all optimized for mobile, tablet, and desktop use.
Objective
To develop a responsive and easy-to-use web application for small business owners and shopkeepers to track, manage, and visualize daily sales data with secure authentication and downloadable reports.
Architecture Overview
Benefits
Secure Access : JWT ensures only authorized access to user-specific sales data.
Visual Insights : Helps users understand trends, top-selling products and performance.
Historical Tracking : Easily review and manage sales history by date or range.
Downloadable Reports : Quickly export sales data for records or
analysis
Device-Friendly: Fully responsive UI makes it convenient for mobile - first users.
Challenges
Implementing real-time and dynamic updates for charts and history.
Managing complex edits and deletions of specific date-based sales entries.
Ensuring fast and correct Excel generation across filter types.
Maintaining robust security while keeping user experience smooth.
Designing a scalable database schema for long-term data retention.
Artificial intelligence
Code Explainer Bot
Description
This project uses Generative AI to build an intelligent assistant that simplifies code comprehension. By integrating GitHub repository parsing with LLM capabilities, the Code Explainer Bot lets users explore files in a project and ask questions like “What does this function do?” or “Can you summarize this code?” The tool delivers natural language explanations, summaries, and suggestions for improvement — perfect for self-learning, code review, or onboarding into new codebases.
Objective
Enable users — especially students, developers, and open-source contributors — to understand unfamiliar codebases by simply providing a GitHub repository link, eliminating the need to manually read and interpret raw code.
Key Features
GitHub repository parsing and code file listing
Code preview with clean formatting
Interactive Q&A on code files (functions, lines, or entire files)
Optional file upload support for local code
Suggestions for improvements and bug detection
File-level and function-level summaries
Tech Stack
OpenAi
Code Llama
AI Model
Code Display & Interaction
Streamlit code block
dropdowns
tabs
Preprocessing & Analysis
Code cleaning
prompt engineering
GitHub Integration
Front End
Streamlit
Architecture Overview
User inputs GitHub repo URL
System fetches and lists code files
User selects file to
explore
Code is displayed in UI
Returns explanations, summaries, suggestions
LLM processes code + question
User asks questions
Benefits
Simplifies code understanding for non-expert users
Speeds up onboarding into open-source or team codebases
Encourages independent learning and collaboration
Reduces cognitive load in code reading and documentation
Assists educators, reviewers, and hackathon participants
Challenges
Accurately parsing and filtering code files from varied repositories
Handling large or complex codebases within token limits
Ensuring explanations are precise and technically sound
Managing multiple languages and coding styles
Providing actionable insights (bugs, improvements) without hallucination
Rewards
01
Winner
₹10 lakhs + Paid Project
02
Winner
₹10 lakhs + Paid Project
03
Winner
₹10 lakhs + Paid Project
Smart Interview Simulator
Description
This project leverages Generative AI to build a smart interview simulator that reads a user's resume and dynamically generates interview questions tailored to their skills, experience, and the selected job role. Users receive real-time feedback, improvement tips, and a performance report — making it an ideal AI interview coach for career preparation.
Key Features
Resume parsing to extract skills, education, and experience
Personalized questiongeneration using LLM
Real-time answer evaluation and scoring
Feedback on strengths and areas of improvement
Performance summary and downloadable report
Optional voice input for answering questions
Objective
Help students and job seekers practice job interviews by providing personalized, AI-driven mock interviews based on their resumes — anytime, anywhere.
Tech Stack
PyMuPDF
PDF
plumber
python-docx
Resume Parsing
Session & Reporting
Streamlit Session
State
PDF plumber
Voice Input
LLM Integration
OpenAi
Front End
Streamlit
Architecture Overview
User uploads resume
Resume parser extracts key information
User selects job
domain
LLM generates personalized interview questions
Feedback and report are generated
AI evaluates and scores the response
User answers via text or voice
Benefits
Makes interview preparation personalized and effective
Enables self-paced practice with real-time feedback
Reduces stress by simulating real interviews
Offers insights on skills and gaps through final report
Encourages iterative improvement and confidence building
Challenges
Parsing unstructured resumes with varying formats
Generating domain-specific, high-quality questions
Evaluating subjective answers with consistent scoring
Managing state and session for multi-turn interactions
Handling multilingual or vague resume inputs
Inventory Management System (CRUD Application using AI)
Objective
Build a lightweight Inventory Management System to allow users to manage products, suppliers, and stock levels in real time.
Description
This CRUD-based app allows small businesses or store owners to track their inventory seamlessly. Users can add, view, update, and delete items, suppliers, and stock levels.
Key Features
Add New Products with attributes like name, category,price, and stock
View Inventory List with filters for category/stock levels
Update Product Info including quantity and status
Delete Products or Suppliers
Supplier Management: Manage vendors, contacts, and purchase history
Tech Stack
Django
Back End
Django ORM
ORM
Database
Supabase
Front End
Streamlit
Architecture Overview
User interacts via Stream lit UI → Adds, updates, or queries product/supplier/stock information in natural language
LangChain processes the user query using an LLM → Converts the natural language prompt into an SQL query or relevant instruction
Generated SQL is executed on the live inventory database → CRUD operation or data retrieval is performed (e.g., stock check, product update)
Results are returned to the backend → Includes raw SQL results (rows, values, messages)
LangChain LLM generates a natural language response → Summarizes the result (e.g., "You have 4 products with low stock.")
Streamlit displays the SQL + natural language result → Shows query for transparency and AI-generated answer for user clarity
Sample Database Schema
id (int)
name (text)
category (text)
quantity (int)
price (float)
supplier_id (foreign key)
Products Table
Suppliers Table
id (int)
name (text)
contact_email (text)
phone (text)
Stock Transactions Table
id
product_id
type (inbound/outbound)
quantity
date
Benefits
Prevents overstocking and understocking
Increases business transparency
Reduces manual effort and paperwork
Enables easy scaling to include more warehouses or products
Challenges
Handling concurrent updates (real-time sync)
Validating supplier-product relationships
Maintaining SQL safety and injection prevention
Maintaining data accuracy during edits or deletes
Technical Requirements
Prevents overstocking and understocking
Increases business transparency
Reduces manual effort and paperwork
Enables easy scaling to include more warehouses or products
Mandatory Requirements
Handling concurrent updates (real-time sync)
Validating supplier-product relationships
Maintaining SQL safety and injection prevention
Maintaining data accuracy during edits or deletes
Registration Opens
31st July 2025
Submit your entries online
Registration Closes
12th August 2025
Last date to Register
Quiz for Shortlisting
14th August 2025
Online quiz for all shortlisted team from registered teams
Final List Announcement
16th August 2025
Top 10 Teams will be announced
Milestone
Date
Description
Hackathon Starts
18th August 2025
After innogration developement starts
Result Day
20th August 2025
Winner team will be announce and Panel discussion starts from 2 pm
Timeline
Timeline
Milestone
Date
Description
Registration Opens
31st July 2025
Submit your entries online
Registration Closes
12th August 2025
Last date to Register
Quiz for Shortlisting
14th August 2025
Online quiz for all shortlisted team from registered teams
Final List Announcement
16th August 2025
Top 10 Teams will be announced
Hackathon Starts
18th August 2025
After innogration developement starts
Result Day
20th August 2025
Winner team will be announce and Panel discussion starts from 2 pm
Registration Closes
12th August 2025
Last date to Register
Rewards
01
Winner
Price will reveal soon.........
02
Winner
Price will reveal soon........
Award Ceremony
On the final day of QHackathon 2025, we’ll honor the brightest minds with:
Exciting Cash Prizes (to be revealed live)
Official Certificates of Excellence
Exclusive Mementos from QHills
Timeline
Milestone
Date
Description
Commencement of Innovation Challenge
Date
Official launch and registration opening
Milestone
Date
Description
Milestone
Date
Description
Milestone
Date
Description
Milestone
Date
Description
Milestone
Date
Description