- Hi, I’m Mario Xerri, an undergraduate computer science student at Cornell University specializing in machine learning systems and software engineering.
- I’m interested in implementing software for mechanical-related systems and solving optimization problems using machine learning and statistical models.
- I'm always open to opportunities for computational-related work in new fields.
- 2D Human Pose Estimation (HPE) Model
- Collaborated with two other students to train and develop a non-pretrained visual transformer deep learning model from scratch for single-person human position estimation using PyTorch
- Minesweeper in OCaml (Spring 2023)
- Collaborated with two other students in the development of a completely function implementation of Minesweeper in OCaml which can be played in the terminal.
- Moves were parsed as commands within the terminal and game states were displayed after each mode
-
Summer 2023 @Brookhaven National Lab:
- Developed and trained a recurrent single object detection convolutional neural network model in PyTorch for event-based camera data utilizing sparse convolution layers from NVIDIA’s Minkowski Engine Library.
- The model performed faster inference compared to an equivalent model using standard convolution when data was over 90% sparse.
- The software was containerized using Podman to enable scalability in the architecture.
-
Summer 2022 @Brookhaven National Lab:
- Worked with a Macromolecular Crystallographer on the justification and testing of a new unit cell representation
- Developed C++ and Python programs to convert unit cells to my mentor’s new cell representation and test database search and structural comparison using this new representation. Tested and debugged my mentor’s code which I was expanding upon.
- Reach me at xmario02@yahoo.com