SENIOR PRODUCT ENGINEER

E-Agent: Learning With AI for Banks and Enterprises

E-Agent: Learning With AI for Banks and Enterprises

IFSS Brings AI Into the Classroom

IFSS trains learners at banks and enterprises, and wants them to work with AI themselves during the session. E-Agent is the web app for that. Each lesson has an assistant, and learners discuss with it to learn faster. The product also includes financial analysis tools.

I worked on this project as Project Manager and Full Stack Developer, and I wrote the backend and the AI part.

E-Agent: Learning With AI for Banks and Enterprises 1

Analyzing Financial Reports From Excel

The hardest part was basic financial report analysis from Excel files. These files have no fixed structure, so they cannot all be read against a single template. The system has to work with whatever content the user provides, not with a predefined layout.

One Assistant per Lesson

Each lesson has its own sub-agent. Each sub-agent carries a system prompt with a different context, so whichever lesson a learner opens, the assistant matches that lesson's content.

Free for Learners, but AI Still Costs Tokens

IFSS provides the platform to learners for free, but every AI reply still costs tokens. Tokens are therefore issued monthly, to limit how much each learner can use.

Tech Stack

  • Next.js
  • TypeScript
  • Firebase