IBM PROJECT
LEARNMATE AI AGENT
Presented By:
Student name :Anjali pathak
College Name :Makhanlal Chaturvedi National University
Department : computer science
OUTLINE
Problem Statement
Technology used
Wow factor
End users
Result
Conclusion
Git-hub Link
Future scope
IBM Certifications
PROBLEM STATEMENT
Many students struggle to find the right learning path among a flood of online courses.
There's no one-size-fits-all roadmap, leading to confusion, wasted time, and incomplete
skill development.
Proposed Solution:
LearnMate is an Agentic AI Coach that personalizes learning pathways using student
interests, current skill level, and time availability. It adapts weekly based on feedback
and progress to ensure a smart, guided learning journey.
TECHNOLOGY USED
IBM cloud lite services
[Link] Studio
IBM Granite 8B Foundation Model
LangGraph Framework
React-based UI
Agent Lab (BETA)
IBM CLOUD SERVICES USED
IBM Cloud Watsonx AI Studio
IBM Cloud Watsonx AI runtime
IBM Cloud Agent Lab
IBM Granite foundation model
WOW FACTORS
Personalized Learning Paths
Learners receive weekly roadmaps customized by interest, skill level, and available time.
Structured as beginner → core → advanced → project-based flow.
Agentic AI Behavior
Acts like a supportive mentor, not just a chatbot.
Uses motivational, user-friendly language with a growth mindset.
Smart Skill Assessment
No tests required—assesses experience level via conversation.
Ensures beginners aren't overwhelmed and advanced users aren't bored.
Interactive Learning Experience
Includes mini-quizzes, hands-on coding tasks, and error-handling tips.
Helps users apply knowledge and reinforce understanding.
Real-Time Adaptation
Roadmap changes based on progress, feedback, or topic switch.
User never feels “stuck” or off-track.
Curated, Trusted Resources
Recommends reliable content (YouTube, docs, platforms).
Avoids overwhelming users with too many or unverified links.
END USERS
College & High School Students
Career Switchers
Educational Institutions
Online Learners in Tech Fields (Frontend, Cybersecurity,
etc.)
RESULTS
LearnMate Agent offers a warm and
engaging welcome, instantly setting
the tone for a personalized learning
experience. With a clear purpose,
adaptive design, and intuitive
interface, it acts as a supportive guide
from the very first interaction—
helping users feel confident,
understood, and ready to begin their
skill-building journey.
RESULTS User Input (Natural Query)
The user asks: “What is the best path
for Frontend development?”
This is a natural language query,
meaning the user can speak freely
without needing structured commands.
2. AI Agent Response
The LearnMate agent interprets the
query and generates a personalized
learning roadmap.
It clearly explains that the roadmap:
Is structured and resource-rich
Assumes a beginner level
RESULTS
RESULTS
FOUNDATION AND TOOLS USED
Foundation Model: granite-3-3-8b-instruct
• The agent is powered by IBM’s Granite 8B Instruct model, a powerful large language
model.
• It enables context-aware responses, dynamic roadmap creation, and human-likeraction
tailored for learners.
• Built for natural language understanding, adaptability, and content generation.
Integrated Tools for Smart Search
• Google Search – Retrieves real-time information from Google.
• DuckDuckGo Search – Offers privacy-first search-based results.
• Wikipedia Search – Extracts topic summaries and definitions from Wikipedia.
• Webcrawler – Collects live data from open websites and pages.
MODULE PARAMETERS:
Why These Settings Are Effective for LearnMate
Balanced Creativity: With a moderate temperature (0.6), the agent gives clear,
logical answers without being too robotic.
Unique Responses: High penalties (1.5–1.9) prevent boring repetition, ensuring
variety in course suggestions and feedback.
Control & Precision: Top P = 0.85 makes outputs both intelligent and reliable—ideal
for education use.
Rich, Informative Output: A token limit of 1997 supports detailed learning paths,
summaries, and suggestions.
CONCLUSION
LearnMate builds personalized learning paths based on user interests and [Link]
ensures every learner gets the right content at the right time.
▪️The AI agent adapts dynamically to feedback and [Link] keeps the journey
flexible, relevant, and motivating.
▪️It acts as a smart mentor—guiding, encouraging, and supporting [Link] stay
confident and engaged throughout the process.
▪️Reduces confusion by recommending only trusted, structured [Link] more
course overload or random content exploration.
▪️Helps users stay on track with weekly plans and learning [Link]
productivity, clarity, and long-term learning outcomes.
GITHUB LINK
GITHUB- [Link]
FUTURE SCOPE
Connect with real LMS like Coursera or YouTube
Dashboard for progress tracking
Voice-enabled interaction for accessibility
Collaborative learning spaces
Skill certificate recommendations after roadmap completion
IBM CERTIFICATIONS
THANK YOU