HUD data product (Ongoing work)

---

The purpose of this project was to address a fundamental issue in the job application process - an incentive system which encourages inaccurate resumes by applicants. Here, we leveraged blockchain databasing to establish a trusted resume authentication system that could solve this problem. The central idea was to enable ex-employers to validate the claims made by an applicant, ensuring that the information provided to the interviewer accurately reflects their past roles and responsibilities.

This was my first attempt at building something at scale. I essentially built a new smart contract on Ethereum. These contracts executed predefined logic, notified ex-employers, confirmed claims, and recorded validated transactions. I developed intuitive interfaces for both job applicants and ex-supervisors, enabling applicants to create their resumes while allowing ex-employers to receive notifications and verify the accuracy of these claims. I spoke to a few universities as early customers.

Tech Stack: Ethereum, Solidity, Truffle, Remix, React.js, Authentication Protocols, NoSQL Database, Data Analytics (Elasticsearch, Kibana).

This project revolved around the application of mathematical tools, including stochastic calculus and differential equations, to model the complex dynamics of asset prices in uncertain financial markets. Primarily, I focused on proving mathematically, and empiraclly, the black-scholes equations. I aimed to create a robust model capable of estimating the fair market value of derivatives according to Delta, Gamma, Theta, Vega, and Rho.

I encoded  processes described by stochastic calculus and relied largely on Ito's lemma. I built a model for derivative pricing, utilizing stochastic differential equations, Ito's Lemma, and Feynman-Kac equations to solve. Risk-neutral pricing theory underpinned my approach, with Brownian motion modeling market fluctuations. To validate our model, we ran Monte-Carlo simulations.

Results are below. Includes some pretty LaTeX math.

https://docs.google.com/document/d/1BA_hFB6HVv8zj-2blZjKoq1NlcyyFcmsQmqpx6G42Gw/edit

A team and I built a user-friendly interactive study-space locator, drawing inspiration from Yelp but with a specified UI for students. We provided users with a convenient tool that would help them find the ideal study environment based on various preferences, such as noise level, amenities like Wi-Fi and printer access, and location convenience. Admittedly, this is a very non-sticky product as once a student finds a good study space, they tend to spend the rest of college there.

Our stack included React.JS for the front-end, Express.JS for the backend, and API interactions for communication with MongoDB. User authentication was implemented for personalized results. Google Maps API enabled a display of study spaces on an interactive map and location-based searching.