University of Pennsylvania

Oberlin College

School Year Abroad - Spain

Belmont Hill School

PhD student in Bioengineering at The University of Pennsylvania

Cedars-Sinai

Research Data Scientist

Department of Computational Biomedicine and Center for Artificial Intelligence.
• Maintained the TPOT Python package for AutoML (over 9,400 stars on GitHub and over 300,000 downloads). Fixed bugs and kept it up-to-date with the latest versions of its dependencies.
•Initiated and led the development of TPOT2, an open-source Python package that uses evolutionary algorithms to automate the construction and tuning of machine learning pipelines. Six researchers in our group actively use this tool in their research on topics such as fairness in AI, genetics, data cleaning, evolutionary algorithms, and AutoML. I presented this project at the Genetic Programming Theory And Practice 2023.
•Significantly improved workflow and feature development of TPOT2 by developing a modular design pattern, creating a prioritized roadmap of tasks (bugs, optimizations, features), suggesting task assignments based on experience and interest, and developing parallelized benchmarking scripts. Working with three developers and two postdocs on maintaining and improving the package.
•Conducting independent research on evolutionary algorithms to optimize TPOT2 to generate interpretable models via improved feature selection and feature engineering.
•Collaborated with scientists to enhance their research projects by applying data science techniques and conducting initial exploratory analysis.
•Built predictive machine learning models for Amyotrophic Lateral Sclerosis (ALS) severity and predicting major adverse cardiovascular events (MACE).
•Used PyTorch deep learning models to segment and detect gallstones using limited imaging data.
•Mentored a Master student intern in designing and executing projects on automated data cleaning and feature engineering.
•Wrote a comprehensive 22-page document for effectively using our High-Performance Computing cluster, including SLURM, SGE, Linux/Bash, Python, Jupyter notebooks, and VSCode guides.
•Tested, benchmarked, and evaluated bioinformatics software.

Programmer / Analyst

Conducted graduate research under Dr. Konrad Kording. 

•	Developed Pytorch-based deep learning architectures to build predictive models of behavior from neural recordings. Used the models to analyze the representation of behaviors in the brain.
•	Investigated the patterns of activity of the neurons in the Frontal Eye Field using MATLAB. 
Work culminated in a Master's Thesis.

Other activities

Penn Biotech Group Healthcare Consulting:
•	Provided consulting services to a large pharmaceutical company on a disease management project regarding a gastrointestinal disorder.

Penn Bio Launch: Biotech Fellow:
•	Worked with a multi-disciplinary team during a 10-week workshop to develop a business plan to translate research from a lab into a biotechnology startup.

Graduate Research Assistant

Worked with graduate student Jenelle Feather in Josh McDermott's Laboratory for Computational Audition as part of the MSRP - Bio summer program at MIT


•	Developed parallelized Python scripts to extract and preprocess audio from YouTube to create a large 3TB data set of natural sounds (labeled by Google AudioSet). 
•	Built an HDF5 database to store labeled audio clips that will be used broadly by the McDermott lab for deep learning and neuroscience research.
•	Trained a Tensorflow convolutional deep learning model on a multiclass classification task using the 3TB Google Audioset to identify which of 527 possible natural sound categories were present in audio clips, achieving a 90\% top-15 accuracy.
•	Work was done on a Linux-based high performance compute (HPC) cluster.
•	Presented at ABRCMS 2017, and the Oberlin Celebration of Undergraduate Research 2017.

Massachusetts Institute of Technology

Research Intern

Worked under Ayoub Daliri in Frank Guenther's SpeechLab during the 2016 USPIN summer program.

•	Implemented a MATLAB-based algorithm to quantitatively assess the biological plausibility and accuracy of computational models of full brain networks through comparison to functional neuroimaging data. Work was done on a high-performance computing cluster.
•	Presented at ABRCMS 2016 (ABRCMS Travel Award) and the Oberlin Celebration of Undergraduate Research 2016.

Boston University

Worked in Leslie Kwakye's multisensory neuroscience lab at Oberlin College.
•	Designed, developed, and ran participants on a Python based virtual reality attentional cueing task for the Oculus Rift that sought to test whether a realistic, immersive environment would lead to enhanced behavioral responses. Animated stock models and built a 3D forest environment using 3ds Max. 
•	Performed statistical analysis of data which showed stronger responses to visual cues confirming conclusions from other studies that the brain prioritizes visual to auditory cues in spatial location.

Research Assistant

Worked in Dr. Paul W Davenport's Physiology lab at the University of Florida

•	Investigated the activity of central pattern generator neurons involved in breathing and swallowing located in the medulla of a cat. This research could help find treatments for people with difficulty swallowing, dysphagia, due to damage sometimes caused by stroke to this area of the brain.
•	Used Spike2 to perform waveform analysis on extracellular recordings of a cat during breathing and induced swallowing. Sorted the results based on patterns of activity and analyzed the results.

Pedro Henrique Ribeiro