Ai Product Software Engineering

Vishwa Pandian

-424-***-****

-*****************@*****.***

Education:

-B.S. Software Engineering, The University of Texas at Dallas, December 2024

-M.S. Computer Science, The University of Texas at Dallas, May 2027 (expected) GitHub: https://github.com/Vishwapandian

SocialAI: https://github.com/Vishwapandian/SocialAI

-Python, Firebase, Google Cloud, Flask

SocialAI-Frontend: https://github.com/Vishwapandian/SocialAI-Frontend

-SwiftUI, Firebase

Matrix: https://github.com/Vishwapandian/Matrix

-Python, React

Story:

I set out to build a useful AI product. My initial hypothesis was simple: 1.We naturally tell our friends about our day.

2.Many people want to journal, but most don't stick with it. From this, I created a conversational journaling AI—a product that gathers user input through natural conversation and automatically compiles daily journal entries. It combined the utility of journaling with the effortlessness of texting a friend. But when I launched it, something unexpected happened: Users weren’t interested in the journal entries.

They loved the conversation itself.

So I asked them a critical question:

"Would you rather talk to this AI than your real human friends?" Most said no.

That pushed me to investigate why.

Functionally, both AI and humans are "text-in, text-out" systems. So what makes talking to a person so different?

I concluded the missing link was emotion.

Large language models, like GPT, simulate the neocortex—the rational, "smart" part of the brain. But what they lack is the limbic system, the emotional core that makes conversations feel alive.

So I built a crude version of a limbic system for AI—a simple emotional engine that gave the AI memory, moods, emotional consistency, and attachment. The difference was night and day. Users found this emotionally aware AI far more engaging than vanilla, Grok, ChatGPT, or other models.

But I didn’t stop there.

The more I observed users, the more I realized:

Humans are passive social creatures.

We evolved not just to interact, but to observe social interactions. Most people prefer watching gameplay over playing, reading a group chat over participating, or following drama over starting it.

That led me to my next hypothesis:

What if we built an "AI Matrix"—a persistent group chat populated by emotionally driven AI characters with their own personalities, social dynamics, and evolving relationships? Now I’m building that:

A dynamic social simulation where humans can observe, join, or influence a living group of AIs.

It’s a superset of the original social AI—an emotionally intelligent, multi-agent system designed to feel as real and messy as human interaction.

This project started as a journaling tool.

It’s becoming a social reality engine.

Contact this candidate