The Fifth Elephant
The Fifth Elephant

Speak at The Fifth Elephant 2026 Annual Conference

Share you work with the community

All submissions

Powered by VideoKen

Swetha A

Swetha A

@swetha_03

Mahita D

Mahita D

@mahita07

From Amnesia to Photographic memory in Agentic Systems

Submitted Jun 25, 2026

I am submitting for: Track 2 - Building & implementing AI tools & agents in production Type of session: Hands-on workshop - 2-4 hours

Objective / Abstract :

You’ve built a few agents, wired them together into a multi-agent system. What comes next? This workshop takes a focused look at one of the most critical and often underestimated parts of an agentic system: memory. Participants will come away with a practical understanding of how memory shapes agent behavior, capabilities, and user experience.

Agenda :

Introduction

A quick grounding in agentic system fundamentals: the core components (LLM, Tools, Planning, Execution, Memory), the agent execution lifecycle (Observe, Reason, Act, Reflect), and why memory is not an afterthought but a first-class concern that directly determines the effectiveness and experience of an agentic system.

Problems with a Stateless Agentic System

What actually breaks when an agent has no state? This section examines the concrete failure modes: loss of context within a conversation, no personalisation, repetitive interactions. It builds the case for why memory matters.

Hands-on Exercise 1:

Build a stateless travel assistant. Run it through various scenarios, observe where it falls short, and identify what is missing.

Short-Term Memory

We cover what short-term memory is and two practical patterns for implementing it:

  • Using workflow state to store user preferences, tool outputs, intermediate decisions, and agent outputs, and how to manage that state across an agent workflow.
  • Passing conversation history in the prompt to maintain continuity and handle follow-up questions effectively.

Managing Conversation History Efficiently

This section covers the decisions that go into managing conversation history well: what must be preserved, what can be dropped, and the common strategies used to keep history useful without letting it become a liability.

Memory Checkpoints

We cover what checkpoints are, why they are needed, and the LangGraph-specific concepts required to implement them: Checkpointers, State persistence, Workflow resumption, and Interrupt and resume patterns.

Hands-on Exercise 2:

Extend the travel assistant to retain details like budget, interests, destinations, and trip duration across a session. The assistant should ask intelligent follow-up questions and maintain coherent context throughout. Run the earlier scenarios/test cases again and compare.

Long-Term Memory

This section covers the limitations of short-term memory, when long-term memory becomes necessary, and the three types: Semantic, Episodic, and Procedural memory.
We then look at architectures for long-term memory: how memory is created, stored, and retrieved. We will also talk about the best practices for maintaining long-term memory in production.

Hands-on Exercise 3:

Extend the travel assistant further. It should now recall preferred destinations, remember budget ranges, make recommendations based on past choices, and pick up conversations where they left off across sessions. The assistant should demonstrate genuine personalisation driven by persistent memory.

Wrap-Up and Discussion

We wrap up with the real trade-offs between memory approaches and common challenges in production followed by a Q & A session.

Key takeaways

By the end of this workshop, participants will:

  • Understand the role of memory in agentic systems
  • Build stateless, session-aware, and persistent agents
  • Implement short-term and long-term memory patterns
  • Use checkpoints effectively in workflow orchestration

Who is this workshop for

  • Developers and engineers building AI-powered applications who want to design reliable, agent-driven workflows
  • ML/AI engineers exploring agentic architectures
  • Tech leads and solution architects designing scalable, multi-agent systems and automation platforms
  • Platform and backend engineers integrating LLMs with real-world systems, APIs, and infrastructure
  • Innovation, R&D, and product teams experimenting with AI agents, orchestration frameworks, and MCP-based extensibility

Start with this talk to understand the fundamentals, then join the workshop for a hands-on deep dive into memory architectures for agentic systems.

https://youtu.be/zWWWi5tfkn4?si=SXbpzCFN62GDcAxa

Speakers’ bio

Swetha A

Swetha is a software developer and GenAI enthusiast working as a Solution Consultant at Sahaj Software. She is passionate about building creative intelligent systems that harness deep learning and generative AI to solve real-world challenges. With hands-on experience in AI-driven projects and a research paper presented at the International Conference on Data Analytics and Management, she has delivered talks on AI agents and facilitated workshops on AI-assisted SDLC, building multi agentic systems empowering people to apply AI thoughtfully.

Linked In: https://www.linkedin.com/in/swetha0302

Past talks & workshops:

Mahita D

Mahita D is a software developer passionate about agentic systems and the evolving role of AI in software development. Her work focuses on exploring innovative ways AI can simplify complex problems and address everyday challenges. Through talks and workshops, she shares her experiences building agentic systems and helping teams turn AI capabilities into practical, real-world solutions.

Linked In: https://www.linkedin.com/in/mahita07

Past talks & workshops:

All submissions

Comments

{{ gettext('Login to leave a comment') }}

{{ gettext('Post a comment…') }}
{{ gettext('New comment') }}
{{ formTitle }}

{{ errorMsg }}

{{ gettext('No comments posted yet') }}

Hosted by

The Fifth Elephant
The Fifth Elephant
Jumpstart better data engineering and AI futures