Salient

Inspiration

Our inspiration for Salient began with the struggle faced by students at UC Berkeley due to the strike of TA's and the growing mismatch of TA's to students. Witnessing the power of chatGPT in our own personal studies and being fascinated by the theory of large language models (LLMs) as simulators capable of emulating any pedagogical mind, we set out to create a platform that harnesses the potential of AI-driven education.

What it does

Salient is an AI-powered learning platform that offers personalized tutoring and course management for students and teachers by combining adaptive lessons, interactive tools, a user-friendly editor, and semantic search. It empowers users to navigate the challenges of acquiring knowledge and track their progress effectively.

How we built it

We employed NextJS for the framework and utilized TailwindCSS along with ShadCN components for styling. Vercel provided deployment and kv cache support, while Supabase served as our vector database (pgvector) and authorization system. Additionally, we integrated Stripe for billing purposes.

Challenges we ran into

As first-time users of the app router and React server components, we experienced numerous hydration issues, which consumed a considerable amount of time and troubleshooting effort.

Accomplishments that we're proud of

We take pride in successfully implementing a wide range of features within a short timeframe to create a robust educational platform.

What we learned

This project has taught us valuable lessons in task prioritization and effective behavior modeling for UX design.

What's next for Salient

Looking ahead, Salient will focus on enhancing support for teachers, improving the notes editor, and exploring the potential of agentic LLM applications for knowledge work automation, ultimately making education more accessible and efficient.

Built With

• nextjs
• stripe
• supabase-(pgvector-db-+-auth)
• tailwindcss-+-shadcn-components-(styling)
• vercel-(deployment-+-kv-cache)