OpportuneAI – Project Story

🌟 Inspiration

Every year, thousands of students in India miss out on scholarships, hackathons, and courses—not because they lack talent, but because the information is scattered, confusing, or hidden behind complex eligibility rules. This challenge is especially severe for learners in tier-2 and tier-3 cities, who often lack the same access to guidance and mentorship as those in metro areas. We built OpportuneAI to bridge this gap by acting as a personal AI mentor.

🧠 What We Learned

Through this project, we explored how to:

Orchestrate multiple AI agents using IBM’s Agent Development Kit (ADK).
Design structured schemas for eligibility and opportunity metadata.
Apply Granite LLMs to simplify eligibility rules and enable multilingual explanations.
Tailor AI solutions to India with a focus on simplicity, inclusivity, and accessibility.

🛠️ How We Built It

Frontend: Streamlit dashboard for student profile input and personalized opportunity display.
Backend: IBM ADK orchestrating specialized agents for analysis, matching, and explanation.
Models: Granite LLMs for rewriting criteria into clear, student-friendly terms (with local language support).
Dataset: Curated CSV/JSON containing sample scholarships, hackathons, and courses.

Workflow:

Student enters profile information.
ProfileAnalyzerAgent parses academic background, skills, and interests.
OpportunityMatcherAgent filters and ranks opportunities from the dataset.
EligibilityExplainerAgent rewrites criteria into plain, understandable language.
Dashboard displays a personalized, easy-to-navigate list of opportunities.

⚡ Challenges We Faced

Extracting and cleaning unstructured data from government/NGO sources.
Standardizing diverse eligibility formats into a unified schema.
Ensuring multilingual support to reach students beyond English speakers.
Integrating Granite with our orchestration pipeline smoothly under time constraints.