EquityLens

Paste your offer. Know what your equity is worth.

Comment

Inspiration

Last semester I co-founded my first startup as part of Northeastern's Husky Startup Challenge. The company was worth $0, but equity division among my two co-founders and I still drove some of our most important decisions. Who had final say, how much to reserve for future employees, and what happens if someone leaves. I learned firsthand that equity is not just a number on paper. It shapes power, trust, and the future of a team.

That experience made me realize how few people actually understand equity, even founders building companies around it. So how overwhelming must it be for someone receiving their first job offer with stock options they have never seen before?

I also watched someone close to me struggle through a long job search as an international student. When he finally got an offer, I wanted to make sure he and everyone like him had the tools to truly understand what they were signing.

EquityLens is for the first-time job seeker who has never heard of a 409A valuation and the international student navigating an unfamiliar system. It translates the jargon into real numbers.

What it does

EquityLens lets you paste any job offer in plain text and instantly get a financial analysis powered by Amazon Nova 2 Lite. It calculates three equity scenarios (Bear, Base, and Bull) based on the company stage and real VC historical data. It also flags your dilution risk, estimates your probability of return, and generates personalized negotiation points.

How I built it

The backend is Python/FastAPI deployed on AWS Lambda via AWS SAM. Amazon Nova 2 Lite with extended thinking powers three agentic endpoints. The first is /api/extract-offer where Nova reads the plain text and extracts structured data like company name, shares, strike price, and vesting schedule. The second is /api/equity-scenarios which calculates Bear/Base/Bull using financial models based on company stage multiples. The third is /api/negotiation-script which generates personalized negotiation points.

The frontend is React/Next.js built with V0 and deployed on Vercel, connecting directly to the AWS Lambda backend via API Gateway.

Challenges I ran into

Getting Nova to extract structured financial data reliably from unformatted offer letters was the hardest part. I also had to solve a Windows/Linux dependency conflict with pywin32 during the SAM build process, and spent significant time debugging the Bear/Base/Bull calculation pipeline between the frontend and backend.

Accomplishments that I am proud of

Building a full-stack AI application from AWS Lambda to a deployed frontend as a Finance student with no prior cloud deployment experience. The Bear/Base/Bull model uses real financial logic, not just hardcoded estimates.

What I learned

How to deploy serverless APIs on AWS, how to use Amazon Nova 2 Lite with extended thinking for structured data extraction, and how agentic AI pipelines work in production. I also learned how to use developer tools I had never touched before: the terminal, Git, Vercel, AWS SAM, and Claude Code. This project taught me that building real software is less about knowing everything upfront and more about learning fast when things break.

What's next for Equitylens

Integrating real-time company data from Crunchbase and PitchBook, adding Levels.fyi compensation benchmarks, and building a cap table simulator for more precise dilution modeling.

Built With

Share this project:

Updates