Skip to content
SSM
← All work

ARCHIVE · ENTRY 04 · CLASS: PUBLIC

Full-stack AI2025

LingualLink

Real-time multilingual chat

A real-time chat platform that translates across languages live, built on a production-shaped stack.

how it resolves
Message
Inference
Translation
WebSocket
Delivery
Problem

Real-time multilingual conversation means low-latency translation, persistent state, and presence — all at once, without the interface feeling laggy.

Built

A FastAPI + Next.js platform with WebSockets for live messaging, Redis for presence and caching, PostgreSQL and MongoDB for state, and optimized ML inference for in-line translation.

Role

Full-stack — backend services, real-time layer, data modelling, and the inference path.

Proof
  • Live cross-language messaging with optimized inference.
  • Production-shaped polyglot persistence (Postgres + MongoDB + Redis).
Stack
FastAPINext.jsWebSocketsRedisPostgreSQLMongoDB
Field note · classified