Inspiration
Our inspiration came to us when thinking of the education sector. We had originally planned to create an AI Student that would sit in a virtual class with other human students. When we wanted to distinguish a motive however, we found that the AI would do more teaching than learning, and knew we didn't want to continue using AI as a student. We realized that the AI could be the teacher instead of pretending to be a student, and thus Lingo was born! The whole concept became the need in the market for a cheap and effective alternative to human speech coaching, which takes far more time and resources than an artificial intelligence teacher does.
What it does
Lingo allows users to grow through targeted questions, exercises, and progress tracking, giving accurate criticisms and critiques of your worst speaking habits. Rushing or dragging, loud or quiet, Lingo helps you fix them all.
How we built it
We built Lingo with a focus on real-time feedback and high-performance AI: Snowflake AI serves as the powerhouse behind our Conversation Mode, orchestrating the dynamic personas and real-time chatting capabilities. Gemma 4 is the analytical brain for our Reading and Verbiage Modes. We utilized its advanced reasoning to parse transcriptions, evaluate pacing and vocal variation, and generate structured, actionable feedback for interview prep. MongoDB acts as our robust database layer, securely storing user authentication, historical exercise analytics, and the live state/evolution level of the user's digital pet. Solana was integrated for our premium tier. We utilized its low-latency blockchain to handle seamless, lightning-fast transactions for users upgrading their accounts to unlock shiny new perks.
Challenges we ran into
One of the biggest hurdles was minimizing latency. When dealing with real-time audio transcriptions and passing that data through large language models like Gemma 4 and Snowflake AI, any delay breaks the illusion of a natural conversation. We had to heavily optimize our API calls and data flow to ensure the real-time tooltips.
Accomplishments that we're proud of
We are incredibly proud of the core gamification loop. Building the dynamic UI that shifts themes based on your pet, complete with smooth Light/Dark mode transitions, makes the app feel polished and alive.
What we learned
Building Lingo taught us a massive amount about AI orchestration. We learned how to prompt and fine-tune Gemma 4 not just to generate text, but to act as a strict, analytical evaluator of human speech patterns. We also gained valuable experience bridging the gap between traditional database structures in MongoDB and decentralized wallet transactions via Solana.
What's next for Lingo
We want to take Lingo's gamification to the next level. In the future, we plan to introduce multiplayer elements, like speech-battle leaderboards and peer-to-peer interview practice. We also want to expand the digital pet ecosystem with more complex evolution trees and a mobile app to let users practice their pitches on the go.
Log in or sign up for Devpost to join the conversation.