ProjectsMay 1, 2024

SpeakEasy App

Overview

SpeakEasy is a real-time voice translation app that lets users speak in one language and hear or read the translation in another—supporting 100+ languages. Built as a cross-platform mobile experience, it targets travelers, language learners, and anyone who needs instant spoken communication across language barriers. The app achieves 95% speech recognition accuracy and 98% translation accuracy in testing, with a polished UI that includes dark and light theme support for comfortable use in any environment.

Tech Stack

The mobile client is built with React Native and Expo, providing a single codebase for iOS and Android. TypeScript enforces type safety across the app layer. A Node.js backend handles API routing and session management, while a Python service processes speech-to-text and machine translation requests. Audio capture, streaming, and playback are orchestrated end-to-end for low-latency translation.

Highlights

Achieved 95% speech recognition accuracy across varied accents and speaking speeds.
Supported 100+ languages for both input recognition and output translation.
Built a cross-platform mobile app with Expo, including dark/light theme switching.
Designed a real-time audio pipeline from microphone capture to translated speech output.

More projects

Python

Deep Researcher RAG System

Multi-agent research pipeline with FAISS vector store, Streamlit UI, and FastAPI backend. Generates comprehensive research reports from a single query using RAG architecture.

Case study View project

Next.js

IdeaClinic Forum

Full-featured forum platform built for DEI, SRMIST using Next.js and Supabase for real-time data and authentication.

Case study View project

TensorFlow

Video Action Recognition

CNN-based action recognition model using MovieNet architecture for Samsung Prism. Achieved 97% training accuracy on compressed video datasets.

Case study