Welcome to my website. Thanks for stopping by. 👋

About Me

Get to know me!

Hi, I'm Melvin He, from Massachusetts! I enjoy solving challenging problems, and I do it with the use of code! I'm a junior at Brown University, where I'm studying Computer Science. In particular, I love working with AI and Systems, and I do a bunch of Full-Stack development too.

Besides coding, some of my other passions are the following:

Brown CSA : I'm a current Family Head for Brown's Chinese-American Students Association where we hold events and panels to introduce Chinese culture to members of the Brown community!

Sports: Whether it's watching or playing, you can catch me cheering on the Celtics and Boston teams, or playing a quick pick-up game of football, tennis, or basketball!

Hobbies: Some of my other hobbies include scenic biking & hiking, playing Super Smash Bros & board games with friends, and learning a new recipe/dish to cook (and eat)!

Skills

GoLang

C/C++

TypeScript

Python

Java

HTML/CSS

Angular

React

Three.js

Express.js

Node.js

MongoDB

SQL

AWS

Azure

Firebase

Postman

Terraform

Jenkins

Docker

Kubernetes

Experiences

Fidelity Investments

ThermoFisher Scientific

Brown CS

Jara

Hack@Brown

igniteCS

Software Engineer Intern @Fidelity Investments

June 2023 - August 2023

I worked on an agile development team to migrate Fidelity Institutional's JSP-based website to Angular with Sitecore as a CMS, contributing to ETF bricklet design, RESTful API development, Jenkins & Terraform CI/CD pipelines, cloud architecture, and unit/integration testing.

Through rotational roles, I also developed APIs for SwiftUI blockchain app for Fidelity Labs (Fidelity's startup incubator) as well as assisted with quantitative equity trading and UI/UX platform design with Fidelity Capital Markets (Fidelity's innovative trading arm).

Featured Projects

Flight Simulator Pro

Three.js and React airplane model flight simulator with infinite mountainous terrain & crash effects!

Onion Routing ACN

Research paper and accompanying Go-based implementation of onion-layered ACN system with end-to-end AES encryption and Diffie-Hellman key exchange, deployable on Docker machines.

IP-TCP

RFC-compliant Internet Protocol and Transmission Control Protocol with circular buffers, retransmission queues, and connection teardown. Built from scratch in GoLang, Lua, and Python.

Jara Unit

IoT-enabled, crank & solar-powered educational device built for children globally in low-income regions like Nepal. Won Develop For Good 2023 Summit's Most Creative award & Most Impactful award.

Realtime Raytracer

Realtime OpenGL graphics pipeline and raytracing engine in C++, built with OOP design to handle complex scene parsing, phong illumination, reflection, point lights, spot lights, attenuation, texture mapping, FBOs, and camera interactability.

HealthChain

Easy-A x Polkadot Harvard 2023 Hackathon award-winning blockchain healthcare and electronic health record data solution built in Rust, incorporating a smart contract developed using the Gear framework and deployed on the Vara network.

Metaverse FinVR Viz

Created VR financial data story visualizations and conducted user experience research using NASA Task Load Index on Meta Quest 2. Gave presentation and interactive metaverse demos for Brown 2023 Spring Research Symposium and ADOCH.

CogniCents

Equity research and financial data analyzer using AWS Sagemaker, BloombergGPT, Verbwire API, and DistillBERT LLMs to give ticker visualization and sentiment analysis on user-inputted 10-Q reports, stock market trends, and real-time news articles. Submitted to HackPrinceton Spring 2023, winning Financial Hack Finalist award.

Caching I/O

Implemented a single slot cache in C to reduce the number of system calls (i.e., buffered read, write, flush) to optimize database system performance to within 5x of stdio's runtime performance.

Breast Cancer BLR Machine Learning Classifier

Implemented Binary Logistic Regression (BLR) with Tikhonov regularization (L2 norm) and cross-validation to perform classification to predict whether or not a given patient has breast cancer based on health data with over 96% accuracy.

StickySign 🤟

IoT environment using OpenCV, 3D projectors, Python scripts, and Sticky Notes to host an AR arena that gamifies learning American Sign Language. Project submitted to HackHarvard 2022, winning top three prize in Accessibility.