Echo - AI companion game development

Echo AI — AI Character & Interactive Relationship Platform

We are looking for an experienced AI/Backend/Full-Stack Developer to help build Echo, an AI-powered interactive character platform focused on creating persistent, evolving relationships between users and AI characters.

Echo is designed to go beyond a traditional chatbot. Each AI character should have its own personality, traits, goals, memories, emotional state, and relationship with the user. Characters should remember important interactions and adapt their behavior over time.

Project Overview

The current application uses:

React + Vite frontend

Node.js + Express backend

Socket.IO for real-time communication

PostgreSQL for persistent data

An EchoProvider abstraction for AI/LLM integration

A browser-side rules engine for offline fallback

The current dialogue system is primarily deterministic. The next stage is to introduce a production-quality LLM-powered architecture while maintaining reliable server-side control over game state and progression.

Key Responsibilities

The developer will work on:

Integrating an LLM provider through the existing EchoProvider architecture

Designing a scalable AI character/personality system

Implementing long-term memory for users and characters

Building working, episodic, and semantic memory layers

Implementing memory retrieval and relevance/salience mechanisms

Designing relationship progression such as trust, familiarity, and character development

Ensuring progression and important state changes are controlled by the backend

Improving real-time communication and session management

Creating reliable APIs and database structures

Improving the existing architecture for scalability and maintainability

Adding automated tests, validation, logging, and error handling

Preparing the system for future multi-character and multi-user expansion

Memory & AI Direction

The desired architecture should allow Echo to remember important events rather than simply sending the last N messages to an LLM.

Potential architecture includes:

Working Memory → Episodic Memory → Semantic Memory → Relationship Memory → Reflection/Summarization → Relevant Context Retrieval → LLM

We are open to strong architectural proposals, including hybrid rules + LLM systems, retrieval-augmented generation, vector search/pgvector, and multi-agent approaches where appropriate.

Ideal Developer

We are looking for someone with strong experience in:

Node.js / TypeScript

React

PostgreSQL

LLM APIs and AI application development

Prompt/context engineering

RAG and vector databases

Long-term memory systems

WebSocket / Socket.IO

Backend architecture and API design

Automated testing and production deployment

Experience building AI companions, games, character systems, conversational AI, or persistent agent systems is highly desirable.

The developer should be capable of reviewing the existing codebase, identifying architectural weaknesses, proposing improvements, and implementing production-quality solutions rather than simply adding features.