100k Mentor Challenge

A cross-platform mobile application aiming to connect 100,000 underrepresented college students with Fortune 500 mentors, built by a three-engineer team at ProMazo. I served as the primary solutions architect and lead engineer, owning the end-to-end design of both the React Native frontend and the AWS serverless backend, and overseeing all engineering decisions alongside two other engineers.

The platform launched on iOS and Android in November 2020, was covered by Tech & Learning, and has been rated 4.2 stars on the App Store with corporate partners including Best Buy, NBC, Pepsi, CDW, Protiviti, and Neiman Marcus.

Impact

The 100k Mentor Challenge launched during the COVID-19 pandemic, when college students had lost internships and campus recruiting had collapsed. We built the entire MVP of the product within six weeks, and then within a few months of launch, the app crossed 1,000 downloads, with corporate partners across retail, media, and professional services providing mentors. Per the Tech & Learning coverage:

"93% of students and 77% of mentors came from an underrepresented background."

The app continues to operate today, with a current App Store rating of 4.2 / 5.

My Role

As solutions architect and lead engineer, I was responsible for:

Designing the full system architecture from the ground up: cloud infrastructure, API contracts, and data models
Building the cross-platform UI in React Native, targeting both iOS and Android from a single codebase
Architecting the serverless AWS backend and selecting every service in the stack
Overseeing development across the three-engineer team and conducting code review

Application Architecture

The frontend is a React Native application deployed to both the Apple App Store and Google Play Store from a shared codebase. I designed the cross-platform UI to give students and mentors a native-feeling experience on each platform while maintaining a single development surface, which was critical for a small team shipping two apps simultaneously.

The backend runs entirely on AWS serverless infrastructure, with an EC2 instance handling the compute-intensive matching algorithm separately:

Service	Role
API Gateway	REST API layer; routes all mobile client requests to Lambda functions
Lambda	Business logic: profile management, match request handling, session tracking, notifications
DynamoDB	Primary datastore for user profiles, match records, and mentorship session history
S3	Profile media storage (photos, avatars) and static asset hosting
EC2	Matching algorithm: computes ranked mentor/mentee candidate lists from preference vectors

Matching Algorithm

The matching engine runs on a dedicated EC2 instance rather than Lambda because of the memory and compute profile of the ranking workload. Given a student's career goals, industry interests, and background, the algorithm produces a ranked list of 4–5 mentor candidates, and does the same in reverse for mentors. Both sides then rank their options and the system resolves a stable match. Separating this workload from the serverless API layer let us tune the instance independently without impacting API latency.

Tech Stack