Transferly

• 

  Language Change

• 

  College selection

• 

  Main UI

• 

  Resume Fix

• 

  Community Q&A

Inspiration

I'm a community college student at Irvine Valley College, and figuring out the transfer process was genuinely confusing. Between ASSIST.org, articulation agreements, professor reviews, and major requirements, there's a lot of scattered information that students have to piece together on their own. Most people either rely on overloaded counselors or just guess. I wanted to build something that actually guides students through all of it.

What it does

Transferly is an AI academic advisor built specifically for community college students planning to transfer to a 4-year university. It helps students:

• Build a semester-by-semester course plan based on their target major and school
• Match their interests to the right major
• Get professor recommendations based on teaching style and difficulty
• Understand which courses transfer and how they articulate

How we built it

The frontend is built with Next.js and Tailwind CSS, deployed on Vercel. The AI layer uses the Claude API for agent-style interactions, including data fetching, plan generation, and conversational advising. The architecture is designed around a hybrid model where Claude handles reasoning and tool use while a fine-tuned local model handles structured plan generation.

Challenges we ran into

Getting the AI to produce consistent, structured semester plans was harder than expected. Generic LLM outputs needed a lot of guardrails to stay accurate to actual articulation requirements. Balancing response quality with latency was also a real tradeoff we had to think through carefully.

Accomplishments that we're proud of

Building something that actually reflects the real transfer planning process, not just a chatbot wrapper. The major matching and course planning features feel genuinely useful for students who have no idea where to start.

What we learned

Prompt engineering for structured outputs is its own skill. We also learned a lot about fine-tuning tradeoffs and when it makes sense to use a smaller specialized model vs a larger general one.

What's next for Transferly

Expanding the dataset to cover more California community colleges, integrating real ASSIST.org articulation data, and refining the fine-tuned model for more accurate multi-semester plan generation.

Built With

• claude-api
• next.js
• tailwind-css
• typescript