Perry Strahl Email & Phone Number

Most recently, I worked as a summer data science intern with Assurant. In the internship, I worked on building a general time series forecasting framework using RNNs, and with this I successfully forecasted the next 12 months of the CoreLogic residential construction cost index with a 5.17% test MAPE. Last summer, prior to starting the master's, I worked as a data science intern with Civic Knowledge. We worked on developing Career Mechanic, an AI system that reads resumes and computes career trajectories to match candidates with jobs. I evaluated different technologies and implemented a solution for classifying education experiences, a process that involved data cleaning and combining semantic search, traditional text search, NLP and neural nets.Machine Learning - Deep neural networks (feedforward networks, RNNs, transformers), adversarial networks (GAN), random forests, kernel methods, support vector machines (SVM), logistic regression, linear models, k nearest neighbors (KNN), k-means clustering- Deep learning architectures and modeling details (weight initialization, activation functions, data/batch normalization, regularization, SGD, learning rates, back-propagation, optimization) - Natural language processing (NLP), text featurization, feature engineering (cleaning, stop words, n-grams, etc.), word embedding, parsing, text similarity, topic modeling, NER, sentiment analysis, classification, data labeling- Model selection, model tuning, hyperparameter optimization, regularization, model interpretation, feature analysis- Strong background in linear algebra, statistics, mathematical optimization, gradient descentTime Series- EDA: STL decomposition, Holt-Winters exponential smoothing, ACF/PACF, ADF stationarity test, Ljung-Box test- Synthesization: TimeGAN- Forecasting: ARIMA, LSTM, GRUData Engineering - Data acquisition/parsing: Parsing from web, APIs, SQL, data cleaning, processing, scraping- Data visualization, validation, mining, analysis, sampling, outlier/ anomaly detectionSoftware Development - Python, R, debugging, OOP, SQL- PythonML/stats suite (scikit-learn, numpy, pandas, tensorflow, keras, scipy, spacy, hugging face, statsmodels) - Linux - Databricks- Paperspace- Git/Github, Jupyter, Microsoft Office (Word, Excel, PowerPoint), Rmarkdown, LaTeX, common IDEs