Trajectory

Inspiration

Students and young professionals often operate without understanding how their daily habits compound into long-term outcomes. While many tools track behaviors, very few connect those behaviors to probabilistic future scenarios. We built Trajectory to answer one core question: “If I continue living like this, where do I end up?”

What it does

Trajectory is an AI-powered life simulation platform that transforms habits into projected academic, financial, and health outcomes. It runs 1,000 Monte Carlo simulations, computes risk probabilities, and performs sensitivity analysis to identify the most influential behavioral drivers. Finally, it uses a grounded AI layer to interpret those quantitative results into clear, actionable insights.

How I built it

We built Trajectory as a full-stack TypeScript application using React, Vite, Tailwind, and Supabase on the frontend. The quantitative core consists of a custom Monte Carlo engine and a sensitivity analysis module built from scratch. We engineered a separate Node + Express backend to securely integrate OpenAI, injecting structured simulation data into the AI prompt to ensure grounded interpretation.

Challenges I ran into

We faced challenges with secure environment variable management and preventing API key exposure. Debugging proxy configurations, CORS issues, and OpenAI quota errors required careful backend troubleshooting. We also had to refine our prompt structure to ensure the AI interpreted quantitative data rather than generating generic advice.

Accomplishments that I'm proud of

We are proud of building a fully custom probabilistic simulation engine instead of relying on heuristic scoring. Our sensitivity analysis provides interpretable drivers of risk and outcome variance, adding real analytical depth. Most importantly, we successfully architected a secure AI interpretation layer that enhances — rather than replaces — deterministic modeling.

What I learned

We learned how to design and deploy a secure full-stack architecture integrating quantitative modeling with AI interpretation. Building our own backend endpoint taught us practical API design, environment management, and error handling. We also gained experience structuring AI prompts around deterministic data to maintain reliability and performance.

What's next for Trajectory

Next, we plan to introduce multi-scenario comparisons that allow users to test alternative life paths side-by-side. We aim to expand our financial and career modeling with deeper real-world data integration. Ultimately, our goal is to evolve Trajectory into a real-time probabilistic decision engine for academic, financial, and health optimization.

Built With

• node.js
• react
• supabase
• tailwind
• typescript
• vite