Jake Skinner Email & Phone Number

New York, Ny, Us

New York, New York, Us

New York, New York, Us

New York, New York, Us

Chicago, Il, Us

Spring 2021: Cloud Computing: Virtualization Server and GPU Cluster Designed, built, and deployed the hardware and software for a Linux based compute cluster in order to run virtualization servers and ML/Deep Learning workloads at scale in the cloudDesigned and implemented a bespoke Postgres based data warehousing architecture/server to be utilized asynchronously by network serversDeveloped a specialized GPU accelerated server (NVIDIA based) for use in video transcoding and the training of deep learning models/environments and arbitrary data science workflowsWinter 2021: The Data Station ProjectDeveloped a suite of analyzers in Python to detect and classify PII data across unstructured heterogeneous datasets, used for ingestion into meta-data models and large scale data profiling algorithmsUtilized OCR (Tesseract) to extract text from PDF and Image based formats for use in NER models and auto-anonymization workflowsFall 2020: Parallel and Distributed Deep Learning Designed and built a knowledge engineering program in Go to train generalized neural networks in parallelWrote an algorithm to schedule and perform mini-batch and stochastic gradient descent in highly parallel and semi-asynchronous environments, achieving a 3x speedup with no decline in validation performanceSpring 2020Built a contextual text summarization service that summarized long and short form news articles into condensed human readable paragraphsDesigned and built a sequence to sequence deep learning model for abstractive text summarization, employing an LSTM based attententive encoder-decoder architecture to summarize news articles without extractionUtilized Keras and Tensorflow (GPU assisted) to clean, parse, and tokenize 50K news articles and generate customized word embeddings to be used for cosine comparisons and the generation of lexicography coherent summaries

Berlin, De

Developed a companion Apple Watch app (Swift and JavaScript) to monitor, track, and quantify tremors for Parkinson’s patients using Apple’s proprietary Movement Disorder API and native accelerometer readings

San Francisco, California, Us

Designed and deployed an ensemble learning system in python to construct and optimize hex-based delivery zones for 220 markets, achieving Pareto improvements in three consecutive deployments.Constructed a customized grid-search algorithm to cross validated and tune model hyperparameters, preventing redundant simulations on a per market basis and reducing total model runtime by 75%Wrote and productized a multivariate log-linear regression travel time model that utilized Uber’s internal routing and traffic API to predict suboptimal restaurant and delivery locations pairings to prevent degraded food quality

San Francisco, California, Us

Launched and scaled UberEats in midwestern markets growing revenue from $0 to > $100M in 1.5 yearsBuilt the first a courier supply balancing model in SQL/python to quantify imbalances in marketplace efficiency and predict changes in demand and supply, model was productized and scaled across all US marketsDeveloped and implemented time and geo-based dispatch and incentive setting strategy for both driver and bicycle partners across Midwest markets to drive improvements in network efficiency

Dublin 2, Ie

Dublin 2, Ie

https://www.accenture.com/us-en

Dublin 2, Ie

https://www.accenture.com/us-en

Nashville, Tn, Us

Full-time Internship

Chicago, Il, Us

Master'S Degree, Computer Science

Bachelor Of Science (Bs), Human And Organizational Development