Life decisions often involve uncertainty, incomplete information, and emotional pressure. I was inspired by how people frequently ask “What if I choose differently?” yet have no safe way to explore possible outcomes. I wanted to build a tool that doesn’t give advice, but instead helps users simulate futures and understand trade offs before committing to a decision. This idea led to LifeSim AI.

What I Built

LifeSim AI is an interactive decision simulation application that allows users to describe real life choices such as career moves, learning paths, or financial plans and explore multiple possible futures. Instead of a single answer, the system generates several outcome paths (best case, realistic, worst case, and alternatives), each with timelines, risks, and key turning points.

The core feature is the what if interaction, where users can modify constraints (time, money, risk tolerance) and instantly see how outcomes change.

How I Built It

The project is built around the Gemini 3 API , which powers the reasoning and simulation engine. User input is first structured into goals, constraints, and assumptions. Gemini 3 then performs multi step reasoning to generate future scenarios and explain them in clear, structured outputs.

The architecture is intentionally simple:

A lightweight frontend for user input and visualization A backend layer that orchestrates prompts and simulations using Gemini 3 Dynamic re simulation triggered by user what if changes

Challenges & Learnings

One of the main challenges was avoiding generic advice and instead focusing on simulation and explanation . I learned how to design prompts that encourage reasoning over long term outcomes and how to use Gemini 3’s low latency responses to create an interactive experience.

This project taught me how powerful AI can be when used not just to answer questions, but to help people think better about their decisions.

Built With

Share this project:

Updates