Hi, I'm Lorenzo
Computer Science Student at University of Michigan
Passionate about machine learning, software development, and applying technology to solve real-world problems.
Education
University of Michigan
Ann Arbor, MIB.S.E. in Computer Science, Minor in Business Administration
December 2025GPA: 3.80/4.00 | Dean's List: 2022-2025
Relevant Coursework:
Introduction to Natural Language Processing, Applied Parallel Programming with GPUs, Web Systems, Database Management Systems, Introduction to Computer Security, Discrete Math, Data Structures and Algorithms, Introduction to Computer Organization, Foundations of Computer Science, Computational Linear Algebra
My Projects
Cyberwright
A desktop app built with Tauri, Next.js, and Rust, utilizing GPT-4o to identify code vulnerabilities with 85% accuracy, categorize issues, and provide actionable remediation steps.
Course Companion
An LLM-powered application integrating university APIs, Canvas LMS, and Google Calendar API, helping students plan and manage their schedules more efficiently.
Generative AI for Photonic Inverse Design
Research project developing deep learning models including tandem networks, VAEs, and GANs for thin-film optic inverse design.
MotionSurfer
An immersive gaming experience that uses computer vision motion tracking to map player movement to key presses, allowing full-body control instead of traditional touch input.
Fresco
AI-powered food management system that tracks expiration dates, provides optimal storage conditions, and suggests recipes to reduce food waste.
Insta485
A client-side dynamic Instagram clone leveraging JavaScript frameworks to enhance interactivity and responsiveness.
MapReduce
A high-performance multi-process, multi-threaded MapReduce server executing distributed data processing tasks efficiently on a single machine.
Search485
A scalable search engine inspired by Google and Bing, designed to efficiently index and retrieve web content.
Graph Search
A graph search and route tracing system utilizing Breadth-First Search (BFS) and Depth-First Search (DFS) to explore connectivity and shortest paths.
DSA Practice
An implementation of priority queues and templated containers, leveraging inheritance and interface programming for efficient data processing.
HashStruct
A project focused on working with hash tables and managing complex data structures through composition for optimized storage and retrieval.
OptiSolve
An implementation of optimization algorithms, including Traveling Salesperson and Knapsack, for solving computationally intensive decision-making problems.
Research Experience
Generative AI for Photonic Inverse Design
September 2023 - May 2024
This research focused on applying generative AI techniques to solve inverse design problems in photonics, specifically for multilayer thin-film optical coatings. The goal was to design structures that achieve desired optical properties by developing machine learning models that can predict optimal designs.
Research Contributions:
- Developed and benchmarked various deep learning models, including tandem networks, VAEs, and GANs
- Generated over 500 unique test datasets using the transfer-matrix method (TMM)
- Enhanced model robustness by 30% in terms of predictive accuracy
- Identified top-performing models with 25% greater accuracy than previous approaches
Technologies & Methods:
Related publication: DOI: 10.29026/oes.2022.210012
Research Interests
Machine Learning for Scientific Applications
Exploring how ML can accelerate scientific discovery and engineering design processes.
Natural Language Processing
Developing models that can understand and generate human language for various applications.
Computer Vision
Working with image and video data to extract meaningful insights and automate visual tasks.
Work Experience
Machine Learning Intern
Ann Arbor, MIU-M Computer Aided Engineering Network Services
January 2025 - Present- Developing ML pipeline using the OpenAI API and LangChain for precise feature extraction and inference, delivering a 92% accuracy rate in compliance issue detection and processing documents in under 250 ms on average
- Administering AWS infrastructure by managing team permissions via IAM and deploying scalable on-demand processing services (Lambda, S3, Fargate) that handle 30+ grant proposals per day, achieving 300 ms latency per file
- Architecting a robust microservices backend with Docker containerization, integrating a FAISS vector store and a Postgres database to improve document processing throughput by 30% and reduce service response times by 40%
Computer Consultant III
Ann Arbor, MIUniversity of Michigan Information and Technology Services
January 2025 - Present- Assisting faculty, staff, and students with technology in Campus Computing Sites facilities and personal device support.
- Troubleshooting site hardware, re-imaging machines, and performing maintenance on printers and other equipment.
- Executing additional tasks assigned by the Site Manager and CCS Technical Support Group.
Project Lead - ML Flight Price Prediction
Ann Arbor, MIMichigan Data Science Team
January 2025 - Present- Leading a team of 20+ members to develop a flight price prediction system with 85% accuracy.
- Building an interactive web application using Streamlit to enable real-time price forecasts for 1,000+ city-pair markets.
- Conducting structured training sessions on data preprocessing, regression techniques, and model optimization.
LLM Augmentation Engineer
Ann Arbor, MIMichigan Data Science Team
September 2024 - January 2025- Integrated proprietary university APIs, Canvas LMS, and Google Calendar to build a student course companion.
- Implemented RAG with LangChain to enhance response accuracy based on course demands.
- Designed personalized study recommendations using modular workflows for optimized assignment prioritization.
Machine Learning Engineer
Ann Arbor, MIMichigan Data Science Team
September 2023 - September 2024- Developed a deep Q-network (DQN) and Deep Monte Carlo agent using data from Carnegie Mellon's Pluribus agent.
- Optimized model hyperparameters to reduce action-selection latency by 20% and improve real-time decision-making.
- Increased agent win rate by 40% against baseline AI opponents through quantitative performance analysis.
Undergraduate Researcher
Ann Arbor, MIUniversity of Michigan Undergraduate Research Opportunity Program
September 2023 - April 2024- Developed deep learning models (tandem networks, VAEs, GANs) for photonic inverse design.
- Used the Transfer-Matrix Method (TMM) in Python to generate and validate model training data.
- Collaborated with PhD supervisors to apply ML techniques to photovoltaic cell design.
Project Lead & ML/CV Engineer
Ann Arbor, MIMHacks
November 2023- Led a team in developing MotionSurfer, an immersive game using computer vision motion tracking.
- Designed a control system allowing players to control in-game movements via body motion.
- Implemented OpenCV, Mediapipe, and CVZone to track and map player movements to game controls.
Data Analyst
Ann Arbor, MISPARK Electric Racing
September 2022 - August 2023- Researched telemetric concepts to gain a fundamental understanding of telemetry analysis and subsequently implementing them on motorcycle data leading to 5% lap reduction time.
- Created program to extract data from motor controller and battery management system and visualizing it using matplotlib library, improving clarity by 60% among team members.
- Coordinated with team of five on data extraction and visualization by delegating tasks and meeting weekly to measure progress and planning semester deadlines for project milestones.
Platform Intern
Genoa, ItalyRulex
June 2021 - July 2021- Completed the Rulex Workflow Certification program, leading to a deeper understanding about workflow executions, troubleshooting and handling errors, implementing modules, alerts, and runtime variables.
- Implemented algorithms like classification, regression, association, clustering, and optimization leading to an enhanced knowledge of practical business applications such as supply chain forecasting.
- Presented weeklong project of Rulex4 workflow implementation on NBA dataset to senior management, resulting in a 40% model accuracy in predicting NBA victories.
My Skills
Programming Languages
Python
Java
C/C++
SQL
JavaScript
TypeScript
MATLAB
Julia
Swift
Frameworks & Libraries
ReactJS
Flask
Django
Node.js
Express
CUDA
Pandas
Scikit-learn
PyTorch
TensorFlow
NumPy
NLTK
Other Skills
AWS
Git
Docker
Linux
Postgres
Agile
Distributed Systems
Networking
Let's Connect
I'm currently looking for internship opportunities and interesting projects to collaborate on.
Get In Touch