ParallelMe

Inspiration

Every major decision creates uncertainty.

Should you start a startup or take the stable job?
Sleep now or stay up studying?
Move to a new city or stay close to family?

Most tools give advice or static analysis. We wanted to build something different: a way to experience both possible futures created by a decision.

ParallelMe was inspired by the idea that decisions create long-term ripple effects, and understanding those trade-offs can help people make better choices.

What it does

ParallelMe is an AI-powered decision simulator that shows how two choices could evolve into different futures.

A user enters a decision with two options, such as:

Start a startup vs take a stable job
Study notes vs rewatch the lecture
Internship in NYC vs internship in Atlanta

The system then:

Parses the decision into Option A and Option B
Classifies the type of decision (study, career, city, lifestyle)
Loads relevant research or real-world data
Simulates two plausible futures
Visualizes both timelines side-by-side
Shows ripple effects and milestone events
Lets users explore branch scenarios
Allows users to ask questions to their future selves
Helps users decide which path best aligns with their priorities

The goal is not to predict the future perfectly, but to help people understand the trade-offs their decisions create.

How we built it

We built ParallelMe as a full-stack web application using Next.js 14, React, and TypeScript.

On the frontend, we used Tailwind CSS for layout and styling, Framer Motion for smooth animations, and Three.js with React Three Fiber to create interactive 3D worlds that represent each timeline visually.

On the backend, we used Next.js API routes to power the simulation engine. For AI generation, we integrated Groq-hosted Llama models, which produce structured outputs for milestones, timelines, metrics, and reflections.

We also built a decision classification layer so the system can route different decisions to the correct evidence engine. For example:

Study decisions use curated research on learning science and retention
Career decisions use labor-market signals and career-oriented data
Lifestyle decisions use behavioral and productivity heuristics

We added multiple ways to explore the outcome, including:

Parallel 3D world simulations
A ripple timeline of consequences
Branching “what if” events
Future-self reflection prompts
A multiplayer Decision Arena for group input

Challenges we ran into

One of the biggest challenges was getting the AI to produce outcomes that felt specific instead of generic. Early versions sounded too much like broad advice.

We improved this by building a structured simulation pipeline: decision classification, evidence injection, and tightly formatted outputs.

Another major challenge was visualization. Traditional dashboards made the experience feel flat, so we shifted toward interactive parallel worlds and consequence maps to make the futures feel more tangible.

We also had to make sure the system stayed useful across very different kinds of decisions, from study habits to career choices, while still feeling coherent.

Accomplishments that we're proud of

Turning decision-making into an interactive simulation rather than a static recommendation
Combining AI reasoning with structured evidence and visualization
Building parallel timeline worlds that make trade-offs feel intuitive
Creating a system that is both reflective and practical

What we learned

We learned that AI becomes much more useful when it is grounded by structure.

Large language models are great at generating narratives, but the experience improves dramatically when you combine them with evidence layers, decision classification, and interactive visual systems.

We also learned that people connect much more strongly with decisions when they can see the downstream consequences rather than just read advice.