NJ Startup Signals

Project Overview

We built NJ Startup Signals during the Stevens Business + AI Hackathon for the 1435 Capital challenge, ultimately placing second. The app brings public signals about New Jersey companies into a map, ranked list, and sector view. Visitors can filter companies, adjust the weight of different signals, and open a company profile to inspect the underlying records and source links.

My Contributions

• Built the data pipelines. I found and connected public data sources, including SEC Form D filings, research grant records from the NSF, NIH, and SBIR, and USPTO patent records. I implemented the ingestion and processing steps that turn those different sources into company and signal data for the app.
• Set up the database. I configured the Supabase database and the data model for companies, source records, and published signals, so the app could serve traceable results rather than a static demo list.
• Connected the APIs. I built the Next.js backend endpoints that return ranked companies, company details, sector information, and data-source status. The frontend uses those endpoints to update results when visitors change filters and signal weights.

How It Works

1. Collect public records. The backend pipeline imports available funding, grant, and patent data from the external sources.
2. Link records to companies. Source records are processed and stored in Supabase with their provenance, then organized into signals for each company.
3. Serve interactive results. The API scores and ranks companies using the visitor’s selected signal weights. The frontend presents the results on a map and in a list, sorted by score and visualized by sector.

Source coverage varies by company. The app separates live results from clearly labeled sample data, and its AI-generated summaries are a separate part of the team’s work.