Michael Popa Email & Phone Number

Bethesda, Maryland, United States

Worked as a research fellow in the Radiation Oncology Branch (ROB) of the National Cancer Institute, Bethesda Campus. Performed feature extraction, selection and engineering on multi-omic (proteomic, metabolomic) data from patients with pathology proven glioblastoma, a grade 4 brain tumor/glioma. This is in conjunction with research and development of AI models that can assist in revealing potential serum biomarkers, in addition to prediction of tumor characteristics (size, location, morphology, etc.) and patient outcomes. The goal is to integrate and aggregate this data with clinical, imaging, and radiation dosimetry data to enhance interpretation of tumor development + resistance, as well as develop scalable AI algorithms that are feasible to use in a clinical setting for tumor control and enhanced tissue response to radiation therapy.

Houston, Texas, United States

Performed biomedical research using machine learning at the Ray lab of Baylor College of Medicine (BCM) under the supervision of Dr. Christopher Ward and Dr. Russel Ray. The goal was to identify key biomarkers of sudden infant death syndrome (SIDS) using simulated autoresuscitation reflex failures in mice. We focused on predicting whether the autoresuscitation reflex would fail within the next 5-10 minutes based on electrocardiogram (ECG) and plethysmography readings.The work involved a team of six, where we coordinated to develop two ML models that would handle integrated engineered tabular data (w/ 1D CNN + LSTM) and spectral/image data (w/ 2D CNN) derived from ECG and plethysmography readings. My concentration was on data and signal processing, feature extraction, feature selection, dimensionality reduction, feature engineering, as well as controlling for artifacts through normalization, scaling, and denoising. The ML pipeline was automated using Git, allowing for the creation of novel datasets and features for all mice.

Houston, Texas, United States

Utilized Python, R, and MATLAB on several data science and machine learning projects (supervised and unsupervised) involving classification + regression models. This includes linear and logistic regression, KNN, kernel SVM, ensemble methods, and deep-learning (CNNs, stacked autoencoders, RNNs) alongside techniques such as model stacking, resampling, dimensionality reduction, and feature selection + engineering. • Investigating correlation between hippocampal volume and memory processing by simulating appropriate limbic system structures (entorhinal + temporal cortex, DG, CA3, CA1)• Determining the principal factor(s) of synaptic activity based on axonal-dendritic proximity• Finding linguistic biomarkers/distinct speech characteristics for individuals at high risk of developing Parkinson's disease (REM sleep behavior disorder or RBD patients)• Reproducing a neural rise-to-threshold model simulating superior colliculus activity in the midbrain of patients diagnosed with schizophrenia + extending it to those with Parkinson's and Multiple System Atrophy (MSA)

Analyzed and assessed large language model (LLM) performance through iterative testing to optimize its capability to self-learn efficiently and align its output to human preferences. The work conducted on numerous projects involved research of industry-specific knowledge and language patterns/trends to improve the following response metrics: accuracy, relevance, completeness, depth, conciseness, formatting, spelling, grammar, as well as tone and safety.

Bachelor'S Degree, Computational Cognitive Sciences, Neuroscience, And Data Science, Gpa: 3.785

Activities and Societies: Neurotech@Rice, Rice Data to Knowledge Lab/D2K, Omega Psi National Honor Society of Cognitive ScienceMajoring.

Associate Of Science - As, Computer Science, Gpa: 4.0

Activities and Societies: Phi Theta Kappa International Honor SocietyDual-enrolled with Frederick Community College throughout high school.

High School Diploma, Gpa: 4.75

Activities and Societies: SkillsUSA through the Frederick County Career and Technology Center (studying Computer Networking.