Responsible for creating project management and notification system for fire suppression contractors. Designed data schema/backend using PostgreSQL, web application via Django and notification/email system using Celery and Redis. Front end done entirely using Bootstrap (CSS and JS). Fully containerized with Docker and deployed on AWS using ECS.

Developing cloud-based software applications to facilitate various processes within SOCOM. Specifically, the creation of REST/GraphQL microservices, deployed in OpenShift/Kubernetes/EKS. My language of choice is Python and the database of choice is PostgreSQL. Heavy use of Docker, SQL, SQLAlchemy, Flask, Django, AWS services such as S3, Lambda, SQS, SNS, EC2, EKS, ElasticSearch among others. I use Jupyter notebooks for data science/visualization daily tasks.

A 6-month program experimenting with analytics using the Python coding language along with Supervised and Unsupervised Machine Learning models with a specialization in Deep Learning using TensorFlow and Keras.o Machine Learning Approach to Quantum Mechanics: Using gradient boosting for making ground state molecular energy predictions ⌾ Feature engineering: identify molecular representation(s) suitable for machine learning ⌾ High dimensional visualization using principal component analysis (PCA) with an RBF kernel ⌾ Hyper-parameter tuning, primarily to combat overfitting ⌾ Notebook: https://www.kaggle.com/mjmurphy28/predicting-ground-state-energy ⌾ Slide Deck: https://github.com/mkm29/DataScience/blob/master/thinkful/unit/3/capstone/PitchDeck.pdf ⌾ Presidential Topic Modeling ⌾ An in-depth NLP analysis of the speeches delivered by the U.S. Presidents at the state of the union from 1790 to 2019 ⌾ Approaches used consist of Bag of Words, Principle Component Analysis (PCA), lemmatization, POS-tagging, Word Tokenization, Latent Semantic Analysis, Latent Dirichlet Allocation, and automatic summarization using Lex Rank, Luhn Summarizer, LSA and Text Rank ⌾ Kaggle link: https://www.kaggle.com/mjmurphy28/presidential-topic-modelingo Deep Learning for Quantum Mechanics ⌾ Developed tree-based model using XGBoost for predicting enthalpy of atomization, based on previous work (see above) ⌾ Refined molecular representation to enable convolutional neural networks to predict enthalpy of atomization energy ⌾ Kernel PCA for visualization ⌾ Network architecture design and hyper-parameter tuning for 1D and 2D convnets ⌾ Data augmentation ⌾ Notebook: https://www.kaggle.com/mjmurphy28/predicting-atomization-energy-qm7 ⌾ Slide Deck: https://github.com/mkm29/DataScience/blob/master/thinkful/unit/final_capstone/Deep%20Learning%20for%20QM.pdf