We are looking for an experienced full-stack developer to build an MVP of a highly personalized AI agent for our web app. The main goal is to create an AI chat system that remembers users over time. The agent should store past conversations and important information, and use that memory to provide more relevant and personalized responses in future interactions. This is not just a basic chatbot. The key requirement is implementing persistent long-term memory, so the AI can remember user information even after weeks or months of inactivity. For example, if a user shares preferences, personal information, or asks questions over time, the system should store that information and retrieve relevant parts later to improve response quality and personalization. --- Core Functionality Required 1. Chat Interface (Frontend) Build a simple web-based chat interface where users can: • Send messages to the AI • Receive responses • Continue conversations over time • Have conversations saved and loaded The interface should be clean and simple. Mobile-friendly is sufficient. Native iOS/Android apps are not required at this stage. Preferred: React or Next.js --- 2. Backend System Build a backend that: • Receives user messages • Stores conversations in a database • Connects to an LLM API (OpenAI, Claude, or similar) • Retrieves relevant past memories • Sends memory + current message to the AI • Returns the AI response to the frontend Preferred: Python (FastAPI) or Node.js --- 3. Long-Term Memory System (Critical Requirement) This is the most important part of the project. The system must: • Store past conversations and important user information • Convert stored information into embeddings • Store embeddings in a vector database • Retrieve relevant memories based on new user messages • Use retrieved memory to improve AI responses Memory must persist across sessions and remain available long-term. Example behavior: User says: “I live in Zurich” Later asks: “What activities do you recommend?” The system should retrieve the location memory and allow the AI to give a personalized response. --- 4. Database and Storage The system should include: • Database to store users, conversations, and messages • Vector database to store embeddings and memory • Proper structure to allow reliable retrieval Preferred options include: • Supabase (preferred for MVP) • Postgres • Pinecone • Weaviate • Chroma You may recommend the best stack. --- 5. AI Integration The system must integrate with an LLM such as: • OpenAI API • Claude API • or similar The backend should handle sending user messages and relevant memory to the model and returning responses. --- Deliverables • Fully working web-based chat MVP • Backend with memory storage and retrieval • Persistent long-term memory functionality • Proper integration with LLM API • Clean and understandable code • Basic documentation of architecture --- You should have experience with: • Building AI or LLM-powered applications • Full-stack web development • Backend development (Python or Node.js) • Database integration • Vector databases and embeddings • OpenAI, Claude, or similar APIs Experience building memory-enabled AI systems is a strong plus. --- To Apply, Please Include 1. Examples of similar projects you have built 2. Your recommended tech stack 3. Estimated timeline for MVP completion 4. Brief explanation of how you would implement long-term memory for this system --- Project Scope This is an MVP project. The goal is to build a working foundation that we can improve and scale later. There is potential for continued collaboration after the MVP is complete.