Skip to Main Content
Talk Beginner First Talk

Doing Multiple Things Safely: Using RPC for Transactions

Proposal status is Rejected
Session Description

How do you ensure data stays correct when one user action triggers multiple steps, like placing an order and updating stock? We ran into issues where partial failures left the system in an inconsistent state.

This talk covers how we solved this by moving critical logic to the server using RPC and transactions.

We’ll share:

  • Why handling multi-step operations on the client can fail

  • Using RPC to combine multiple actions into a single safe operation

  • Ensuring consistency with transactions (all succeed or none do)

  • Handling errors and edge cases in real-world scenarios

  • When to use RPC vs regular API calls

By the end, you’ll understand how to design safer backend operations and avoid inconsistent data in your applications.

References

Session Categories

Engineering practice - productivity, debugging
Other

Which track are you applying for?

Main track

Speakers

Hrithik Gavankar
Hrithik Gavankar Software Engineer | Red Hat

I am a Software Engineer at Red Hat working on Ansible developer tools, AI agents, and Model Context Protocol (MCP) systems. As an open source contributor, I'm passionate about building intelligent developer experiences, advancing automation, and enabling developers through modern DevTools and AI. I enjoy sharing practical insights on open source, AI-assisted development, and emerging technologies.

Gargi Kshirsagar
Gargi Kshirsagar Associate Software Engineer | IBM

I am a Software Developer at IBM specializing in data engineering, Snowflake, Workato automation, Python, Streamlit, and agentic AI development. I’m passionate about building intelligent, scalable solutions and enjoy sharing practical insights on data platforms, automation, and AI-driven development.