My name is Junyang Lu, senior at UMass Amherst studying Computer Science and Mathematics. Currently, I am interning for Fidelity Investment and am part of the Minutemen Alternative Investment Fund as a quant trader. I aspire to explore the possibilities into Quant and Machine Learning. Also check out my Devpost here.

Lose My Mind (feat. Doja Cat) [From F1 The Movie]

Don Toliver, Doja Cat, F1 The Album

▶ Play Now

• Binary Classification Neural Network
  • Built a neural network from scrach using nothing but Numpy
  • Functioning forward/backward propagation and gradient descent
• What LLM Can I Run?
  • Cross-platform CLI tool to determine the best stuiable LLM to run based on user hardware and available RAM/VRAM
• Fine-tuned Debator
  • Fine-tuned SmolLM2 and Qwen-2 on scraped Reddit data across 20 controversial social issues
  • Evaluated performance using BERTScore
• Flame Detection
  • A convolutional neural network to detect forest fires

• Factor Model Based Portfolio Optimization
  • Focused on volatility and beta calculations
• Sentiment-Oriented Trading Algorithm
  • Fine-tunning OPT-2.7B on news headlines to determine buy/sell on targeted stock
  • Inspired by this paper

• Asciimov
  • Rust-based CLI tool to convert images into ascii art in the terminal

• Gitcoin
  • Crypotcurrency based on Git commit hashes with functioning CLI for mining and transferring
  • Now with vritual and physical copies!
• UMass Rocket Team
  • Programmed microcontrollers to collect orientation and acceleration data during rocket ascend
  • Configured and wired LEDs on payload drone

• Personalization of Large Language Models in Context of Education (ML4ED)
  • Building a full-stack application to collect teacher-student interaction data for future LLM training/fine-tuning
• Random Walk & Quantum Walk
  • Simulated random/quantum walk on grid/cycle graphs with different intial conditions and investigated the impacts on the resulting entropy
  • Conference poster