Bence Lukacsy
Barrington, RI | (415) 638-7304 | bence.luk[email protected] | linkedin.com/in/blukacsy | github.com/senseibence
Education
Stony Brook University Stony Brook, NY
Bachelor of Science in Computer Science and Mathematics Aug. 2023 – May 2027
•
GPA: 4.0/4.0
•
Honors College and Honors in Computer Science
•
Relevant Coursework: Graduate Parallel Computing (AMS 530), Honors Algorithms (CSE 385), Probability
Theory (AMS 311), Honors Data Structures (CSE 260)
Experience
Machine Learning Research Assistant Sept. 2024 – Present
McKinnon Rosati Laboratory Stony Brook, NY
•
Led a 4-person team to develop a supervised deep learning pipeline for automatic cell-type classification, achieving
93% accuracy
•
Optimized JAX model training by leveraging CUDA and Nvidia A100 GPUs on the SeaWulf HPC cluster, scaling
to ∼4 million cells
•
Presented research at the SBU Undergraduate Research Symposium, applying results to multiple scRNA-seq
datasets
Population Genetics Research Assistant Sept. 2024 – Present
True Laboratory Stony Brook, NY
•
Designed a multiprocessed, individual-based Monte Carlo simulation in Python and R to model asexual
reproduction in fruit flies
•
Used the SeaWulf HPC cluster and Slurm job arrays to scale simulations to 96 CPUs and 100+ parameter sets
•
Developed a high-performance simulation framework to explore parameters influencing asexual reproduction in very
large population sizes
Instructor May 2024 – Aug. 2024
iD Tech Remote
•
Led three week-long machine learning courses for high school students; rated highest among ML instructors
•
Provided personalized tutoring in programming, algebra, and chess to middle and high school students
Projects
Content Moderation Bot for Discord | JavaScript, discord.js, MongoDB, PM2 June 2022 – Present
•
Created a Discord bot that moderates content for 150,000+ users across 500+ online communities
•
Implemented a robust regex filter that detects and deletes messages containing profanity
•
Utilizing MongoDB to store state data such as on/off status for moderation features
•
Self-hosting using Raspberry Pi 5 with PM2, achieving over 99% uptime
Colorectal Cancer Risk Assessment | Python, Flask, XGBoost, TensorFlow Oct. 2022 – May 2023
•
Trained a deep learning model using data from NCI’s Cancer Data Access System
•
Developed a full-stack web application and REST API with Python Flask and JavaScript
•
Initially deployed using AWS Elastic Beanstalk, now self-hosted with Nginx on Raspberry Pi 4
•
Regeneron ISEF finalist and 1st place winner at RISEF
Crypto Arbitrage Bot | JavaScript, Harmony-JS, Puppeteer, Nodemailer Sept. 2022 – Jan. 2023
•
Built a crypto token price tracker that uses Nodemailer to send alerts based on price signals
•
Developed a web scraper using Puppeteer to identify and profit from mispricings on DeFi exchanges
•
Integrated MetaMask into a dApp prototype using the Ethereum provider API and Harmony SDK
Skills
Languages: Python, Java, C/C++, JavaScript, HTML/CSS, MATLAB, Bash, R
Machine Learning: JAX, Keras, TensorFlow, PyTorch, XGBoost, scikit-learn, NumPy, pandas
Technologies: Flask, Node.js, Git, Linux, Slurm, MongoDB, Nginx, AWS, CUDA, MPI, PM2
