Project

Applied Memory for Agents

Hard

38 completions

~ 2 hours

4.0

Learn to build an AI assistant with persistent memory and semantic recall by implementing conversation history strategies, structured key-value storage, vector search, and tool-based task management. By the end, you will have a personal agent that remembers past interactions, retrieves information, and updates tasks through autonomous tool calls.

Provided by

JetBrains Academy

About

AI assistants become far more useful when they can remember things and act on them. In this project, you will build your own personal AI task assistant that maintains conversation context, stores information across sessions, and uses tools to create, search, and update tasks. Instead of just answering questions, your assistant will recall user details, track progress, and help manage daily to-dos — just like a real productivity companion.

Graduate project

This project covers the core topics of the AI Agents: Theory and Practice course, making it sufficiently challenging to be a proud addition to your portfolio.

At least one graduate project is required to complete the course.

What you'll learn

Once you choose a project, we'll provide you with a study plan that includes all the necessary topics from your course to get it built. Here’s what awaits you:

Implement three conversation memory strategies (keep all, sliding window, and summarization) to control how chat history is stored and maintained across interactions.

Implement two persistent memory systems — a key-value StructuredStore with TinyDB and a semantic VectorStore with ChromaDB — enabling agents to store, query, and clear both structured and unstructured information across sessions.

Build a full personal assistant agent that maintains conversation context, stores tasks persistently, and calls tools (create, search, update tasks) automatically.

Reviews

Youssef Boukhadda

4 weeks ago

3.3

not using openai so i had to adapt with genai api. but still fun project for starters

User 619990500

4 weeks ago

3.7

In tool definitions defining key value parameters more descriptively with multiple options for llm will help the resolve any ambiguities from llm outputs

Frank Vreys

1 month ago

3.7

Step 3 was rather difficult. I would say it is more then 2 hours to finish this project.

4.0

Learners who completed this project within the AI Agents: Theory and Practice course rated it as follows:

Usefulness

4.4

Fun

4.0

Clarity

3.7