SCHOLARBOT- RESEARCH AI AGENT

IBM HACKATHON PROJECT

AI-Powered Research Assistant�Using IBM Cloud Lite & IBM Granite Foundation Model

​

Presented By : SATYAM PANDEY

College Name : Thadomal Shahani Engineering College (TSEC)

Department ::Bachelor of Technology (BTECH)

Course: Computer science and engineering

OUTLINE

• Problem Statement
• Technology used
• Wow factor
• End users
• Result
• Conclusion
• Git-hub Link
• Future scope
• IBM Certifications

​

​

PROBLEM STATEMENT

�

❓ What’s the Challenge?

Researchers, students, and professionals often face difficulties like:

📚 Information overload when reviewing academic materials

⏳ Time-consuming manual research and literature review

🔍 Lack of centralized tools for querying across different AI services

🧠 No single intelligent assistant that understands context and evolves with user input

​

💡 Our Solution

AICademic – Watson Research Agent�An AI-powered research assistant that:

🤖 Uses IBM Granite & Watson ML for intelligent query understanding

🔄 Offers multiple backend options: Python, Node.js, Java, Scala

💬 Supports real-time chat + research modes

🎨 Has a sleek React-based UI

🔐 Provides seamless configuration and testing

🌐 Is platform-independent and mobile-friendly

​

TECHNOLOGY USED

​

IBM cloud lite services

Natural Language Processing (NLP)

Retrieval Augmented Generation (RAG)

IBM Granite model

IBM CLOUD SERVICES USED

ResearchAgent/

├── config.env # Environment variables (API keys, Watson config)

│

├── backend/ # All backend implementations

│ ├── python_server.py # Flask server (Python)

│ ├── node_server.js # Express server (Node.js)

│ ├── WatsonResearchAgent.java # Java backend

│ ├── WatsonResearchAgentScala.scala # Scala backend

│ ├── requirements.txt # Python dependencies

│ └── package.json # Node.js dependencies

│

├── frontend/ # React frontend

│ ├── public/

│ │ └── index.html

│ └── src/

│ ├── App.js # Main UI logic

│ ├── index.js # Entry point

│ ├── index.css # Tailwind CSS + styling

│ └── components/ # Future component separation

│

├── scripts/ # Dev tools and setup scripts

• IBM Cloud Watsonx AI Studio
• IBM Cloud Watsonx AI runtime
• IBM Cloud Agent Lab
• IBM Granite foundation model

​

WOW FACTORS

🔁 Multi-Backend Architecture

🐍 Python | 🟢 Node.js | ☕ Java | 🟡 Scala — One frontend, many brains

• Unique approach that supports multiple backend engines
• Flexibility for enterprise-level integration and scalability

💬 Dual Interaction Modes

Simple Q&A for quick queries + Chat Mode for conversations

• Smart chat retains history and context
• Gives a realistic research assistant feel

🌐 React Frontend + Real-time API Switching Choose your backend in real time, test connection instantly

• Live connection
• CheckFully responsive UI (desktop + mobile)
• Clean UX with modern design system (Tailwind / Vite)

🧠 Powered by IBM Granite + RAG

Combine generation + retrieval = smarter answers

• Uses Granite Foundation Model for advanced reasoning
• Integrated Retrieval-Augmented Generation (RAG) with PDF/document knowledge
• Provides context-aware, research-grade answers

⚙️ Zero Vendor Lock-in

Can run on IBM Cloud Lite, or locally — fully open & extensible

• IBM Cloud APIs, but dev-friendly for local testing
• Configurable through .env and CLI

🚀 Bonus WOW:

• Fully documented PowerShell scripts to spin up each backend
• End-to-end RESTful APIs with health checks and custom routing
• Designed with developer experience (DX) in mind

​

END USERS

🎓 Students

• Summarize academic papers, generate assignments, get quick topic overviews
• Perfect for research projects, thesis writing, and exam prep

​

🧑‍🏫 Researchers & Professors

• Automate literature reviews
• Draft abstracts, related work, and research summaries
• Validate and cross-reference research findings

​

🧑‍🔬 Data Scientists & AI Engineers

• Use as a knowledge base for AI papers, algorithms, and trends
• Explore cross-domain use cases using RAG

​

📚 Academic Institutions & Libraries

• Provide AI research assistant kiosks or portals
• Help students with queries on-demand
• Can be deployed across multiple departments

​

🧠 Content Writers & EdTech Startups

• Generate education-oriented content faster
• Integrate backend as an API for learning platforms

​

RESULTS

RESULTS

RESULTS

Deployed AI Agent

CONCLUSION

🎯 What We Achieved

• Built an intelligent AI-powered research assistant using IBM Watsonx.ai
• Leveraged Granite foundation model with RAG and NLP to simulate a real academic assistant
• Developed a multi-backend architecture (Python, Node.js, Java, Scala) with a sleek React frontend
• Enabled dual interaction modes: Simple Research and Chat Mode
• Successfully deployed and tested using IBM Cloud Lite services

​

🚀 Why It Matters

• Saves time, boosts productivity, and enhances the research workflow
• Scalable for real-world academic use cases
• Flexible architecture for developers and institutions alike

​

💬 Final Thoughts

"We didn't just build a chatbot — we built a future-ready, open-source Research Intelligence Platform powered by IBM's AI ecosystem."

GITHUB LINK

🔗 GitHub Repository

🚀 Explore the full source code, backend logic, frontend UI, setup scripts, and IBM Cloud integration

here:👉 ResearchAgent

​

💡 Star it. Fork it. Use it.

• This project is open-source and built to help students, researchers, and developers build AI-powered tools with IBM Watsonx and Granite.
• Let’s reimagine academic research—powered by AI. 🌐🤖📚

FUTURE SCOPE

🧠 1. Voice-Based Research Assistant

Integrate speech-to-text and text-to-speech for a hands-free academic assistant

• Helpful for visually impaired users and multitaskers
• Enables smart assistants on mobile & voice platforms

📚 2. PDF Upload + Summarization

Let users upload full research papers and receive:

• Key takeaways
• Summaries per section
• Highlight extraction with RAG

🌐 3. Multilingual Support

Expand usability across regional and global languages using:

• IBM Watson Language Translator
• Granite’s multilingual understanding

💾 4. User Profiles + History

• Enable login, personalization, and saved research threads
• Track progress, bookmarks, and export data
• Provide tailored suggestions based on history

🤝 5. Integration with Learning Platforms

• Turn the agent into a plugin or API for:
• University portals,LMS systems (like Moodle, Google Classroom),EdTech apps

IBM CERTIFICATIONS

• Screenshot/ credly certificate

( Getting started with AI )

IBM CERTIFICATIONS

• Screenshot/ credly certificate

(Journey to Cloud: Envisioning Your Solution)

• Screenshot of certificate

(Lab: Retrieval Augmented Generation with LangChain)

IBM Certifications

THANK YOU