Matvey (Matthew) Moisseyev Matvey (Matthew) Moisseyev

I am an undergraduate majoring in mathematics, data science and computer science at the Raikes School under University of Nebraska - Lincoln. I conduct research in applied math under Dr. Hongfeng Yu and Dr. Huijing Du. I’m also an intern at Nelnet developing internal software and performing quantitative market research. My overall interests lie in math, machine learning and finance.

Matvey (Matthew) Moisseyev

Research

Scalable Multi-GPU Simulation of 3D Multicellular Growth with RNN-Based Workload Balancing

Date: August 21, 2026; Authors: Matvey Moisseyev, Huijing Du, Dandan Zheng, Chi Zhang, Hongfeng Yu
Paper Code BibTex TL;DR
Scalable Multi-GPU Simulation of 3D Multicellular Growth with RNN-Based Workload Balancing

Acoustic Streaming of a Viscoelastic Fluid in a Wall-Actuated Microchannel

Date: January 13, 2026; Authors: Matthew Moisseyev
Poster TL;DR
Acoustic Streaming of a Viscoelastic Fluid in a Wall-Actuated Microchannel

Modeling Electrolyte Anions at the BiVO₄/Water Interface: Thermostat Validation and Hole-Polaron SO₄²⁻ Adsorption Energetics

Date: May 12, 2025; Authors: Matthew Moisseyev
Poster TL;DR
Modeling Electrolyte Anions at the BiVO₄/Water Interface: Thermostat Validation and Hole-Polaron SO₄²⁻ Adsorption Energetics

Projects

MLP Neural Network from Scratch

Tools: C, OpenCL

Implemented forward propagation and hand-derived backpropagation in pure C with no ML libraries. Wrote custom OpenCL kernels for matrix operations (3x speedup), achieving ~98% accuracy on MNIST.

Code Demo
MLP Neural Network from Scratch image

Smoothed Particle Hydrodynamics

Tools: C, CUDA, Python

A simple project for learning fluid dynamics and employing CUDA for physical simulations. Uses smooth particle hydrodynamics approach to model fluids. Implemented uniform-grid spatial binning for fast neighbor lookups, reducing the time complexity to near linear and scaling the simulation to roughly one million particles.

Code Demo
Smoothed Particle Hydrodynamics image

Hackathons

Lockey

An AI-powered corporate travel concierge built for the modern enterprise. Lockey combines Google Gemini 1.5 Pro, MongoDB Atlas Vector Search, and a real-time crisis engine to guide business travelers through the entire trip lifecycle—surfacing policy violations before they happen, drafting manager approval emails, and automatically rebooking disrupted flights, all through a conversational 2D mascot.

Code

Sanos

Sanos monitors your entire Next.js stack in real time, traces each error back to the exact commit that introduced it, and ships the fix as a reviewed pull request... automatically. Your team wakes up to an already-diagnosed bug and a ready-to-merge solution instead of a 2am page.

Code

knkt

A Bluetooth-powered, AI-driven mobile platform built by students, for students. knkt passively scans for nearby peers, computes cosine similarity across skill embeddings and focus areas, and delivers a personalized match summary the moment two students with complementary profiles are in the same vicinity — turning physical proximity into meaningful connections for startups, research, and everything in between.

Code

BeenDone

BeenDone takes your startup idea, searches a vectorized database of YC-backed companies, and returns an AI-powered competitive analysis with feature comparisons, market positioning, and pivot suggestions, all before you write a single line of code. It turns weeks of manual market research into a few seconds of vector search.

Code

Banana Coin

Banana Coin is a simple trading game where you buy and sell coins as the price swings every second, riding the spikes and bailing before the crashes. Leave too long and your wallet rots, so you can unlock minion bots that trade for you while you're away. It's a tongue-in-cheek jab at the anxiety of always-on markets.

Code