Clark Grubb Email & Phone Number

Denver, Colorado, Us

Sherman Oaks, California, Us

Built feature extraction component of text classification system.Built infrastructure for training models asynchronously and on-the-fly from a web service.Refactored back-end services. Goals were micro-services architecture and consistent API. Services I rebuilt: sentiment analysis and NER (Python webserver around a third-party C++ library), geolocation (Node.js), ethnicity detection (Node.js), psychological profiling (Node.js), url lookups (Node.js).Implemented unsupervised learning techniques for data exploration: k-means, hierarchical clustering, principal component analysis, t-SNE, word embeddings.De-duped text using w-shingling (aka character n-grams).Implemented mini-languages for Elasticsearch filters and for notification conditions.Built backend for notification system.Implemented Spark jobs which ran locally or on EMR using S3. Jobs counted n-grams (character or word), performed twitter specific tokenization, found documents containing keywords.Developed in Node.js: maintenance tasks and adding features to the anomaly detection system and the API used by the front end web app.Packaged services in Docker containers and made them deployable via ansible.Analyzed time series for classification (anomaly detection, cyclical pattern detection, upward and downward trend detection) and forecasting (using ARIMA and Holt Winters).Built system for using Mechanical Turk API to collect labeled data for text classification.Built experimental service with Kakfa, Spark Streaming, and Drools.LANGUAGES and LIBRARIES: • Python (flask, pyparsing, celery, numpy, scipy, scikit-learn, gensim, nltk, spacy, rpy2), • Node.js (PEG.js)• Scala (spark)• R (forecast)OTHER INFRASTRUCTURE:• Elasticsearch, MongoDB, Redis, S3, Cassanda• Linux (Ubuntu, Alpine), AWS, RabbitMQ, Git, Ansible, Docker, make, bash

Santa Monica, Ca, Us

Evaluated revenue prediction and text de-duplication algorithms with Python, R, Elasticsearch. Improved performance of text de-duplication algorithms with C++ and Parallel Make. Evaluated competition with Nutch, Ruby, R.Implemented real time back end data service with Python, Redis, and RabbitMQ.Processed web analytics data with Python, Pig, PostgreSQL, Java, Hadoop, and Netezza.Web development with PHP, JavaScript, MongoDB, Python. Google Analytics, Google Tag Manager, sitemap generation.Linux, Git, Mercurial.

Mobile app development using Python, JavaScript, CSS, and XSLT. Linux, Git.

Dallas, Tx, Us

Assembled business listing data for Yellowpages.com using Ruby and PostgreSQL. Engaged lots of cores by means of make -j and ssh. Developed fuzzy name matching techniques and evaluated them with R. Managed up to 4 engineers. Linux, SVN.

Website development with Ruby, Rails, Flash, JavaScript, PostgreSQL, Apache/Mongrel, and memcached. Wrote backend video processing for duplicate detection in C using ffmpeg and libjpeg. Instrumented the website for Omniture. Database tuning. Generated lots of graphs with MRTG. Linux, SVN.

Seattle, Wa, Us

Fraud detection and third party services using C++, Apache, and Oracle. Customer service applications using C++/C, Perl, Java, and Oracle with CORBA for middleware. Managed Perl developers and tech support specialists. Linux, Digital Unix, Perforce, CVS.

M.A., Mathematics

B.S./B.A., Mathematics/Linguistics