Alby - Learning Buddy

Alby our brain shaped language learning buddy!

Inspiration

America is often called a melting pot, but for many refugees arriving here, the transition is much harder than expected. Through our experience tutoring English to refugees, we saw firsthand how language barriers prevent talented people from reaching their full potential. Master's graduates from Ukraine and Venezuela (engineers, accountants, and lawyers) were stuck in low-skill jobs, simply because they struggled with everyday English conversation.

One refugee once told us that our English lessons were their lifeline to a better job. It wasn't grammar rules or writing that held them back, it was real-world speaking: job interviews, restaurant conversations, doctor visits. Without strong listening and speaking skills, they couldn't fully access the "land of opportunity."

That’s what inspired us to build Alby.

What it does

Alby (Our Adaptive Learning Buddy) helps non-native English speakers practice real-world conversations. It uses OpenAI’s language model and browser speech recognition to listen to the user's English, provide friendly corrections, and keep conversations flowing. Learners can pick a ready-made topic (like a doctor's appointment or job interview) or type in any topic they want to practice. Alby also offers roleplay options, helping users build spontaneous speaking skills critical for real life situations.

How we built it

We built Alby using HTML, CSS, and JavaScript for the front-end. We integrated OpenAI’s GPT-4 API to generate adaptive conversation prompts, corrections, and encouragement. We watched many YouTube tutorials to give us more insight in how to being Alby to life. We added browser-based speech recognition so users could practice speaking, not just typing. The app adaptively adjusts to the user's needs, allowing for both general practice or full roleplay.

Challenges we ran into

None of us had worked with OpenAI’s API or browser speech recognition before. We had to learn quickly, experimenting with prompts, debugging API requests, and designing a responsive UI, while making sure the app stayed focused on solving the real-world problem we care about. It was also tough making the conversations feel natural, not robotic. We went from very little experience to using every online resource we could online to help fill our gaps of knowledge.

Accomplishments that we're proud of

We're proud that we created something meaningful in such a short time that we see could be helpful for our tutoring sessions. Alby isn’t just another app; it’s a real buddy for language learners, especially refugees and immigrants who deserve better opportunities. We stayed committed to building something that focuses on real-world speaking practice, not just memorizing rules. Teamwork was essential in incorporating new ideas.

What we learned

We learned that real language learning happens through conversation, not just textbooks. Working with refugees showed us the gap between knowing grammar and confidently speaking in real-life situations. We also learned a lot about combining AI, speech recognition, and user-centered design under time pressure to make something that matters deeply to us.

What's next for Alby - Learning Buddy

We want to expand Alby to support new languages like Spanish and Arabic. We plan to add new topic packs, including immigration interviews, networking, and social events. Long-term, we aim to make Alby mobile-friendly so users can practice conversations anywhere, anytime, and eventually turn Alby into a learning buddy for any subject, from history to coding. Because real learning solves real problems, and that's exactly what Alby is here to do.