Led the development of the experimentation and modeling platform used by NLP researchers. Evaluated, procured, and deployed Pachyderm for automated data versioning. Deployed and managed the ongoing support of JupyterHub for research experimentation.Was in charge of the evaluation of model serving frameworks. Worked with partner teams, infrastructure, security, etc. to deploy Seldon across the company to support model inference requirements.Delivered LivePerson's "stream processing" production pipeline (Kotlin) enabling 1000 predictions/sec using multilingual BERT based encoders with decoupled classifiers for computational efficiency.Collaborated with NLP researchers to develop an in-house multilingual BERT encoder to support nine languages. The model met functional requirements with respect to accuracy and performance on classification tasks in addition to non-functional requirements such as computational load and latency requirements for other services.

Applied mathematics and computational science to problems in remote sensing and earth science. Developed methods and technology to give scientists and their collaborators better tools for information exploration, visualization, data integration and compression.

Work directly with the VP of Analytics & Data and the CTO to assist in designing, developing, and implementing the data management and business intelligence platform that enables our data-driven decision-making.Responsibilities:- Managing, growing and improving on our portfolio of reports, dashboards and ongoing data processing jobs- Working hands-on with the large amounts of data we collect to transport, transform and digest it in whatever way necessary to find the key learnings that would otherwise go unnoticed- Playing a critical role in the architecture and design of our database as we continue to expand our application and increase volume- Identifying points of inefficiency in the collection/processing of data for both reporting and in our application

Executed existing traffic projects for various clients (placing pixels, tagging, reporting and analysis); Initiated activities for existing and new clients; Created a collaboration with analysts and Channel Managers; Set up and implemented automation procedures;

Working on building a Django based web platform that integrates (via RESTful APIs) with some of the AdTech industries' leading ad exchanges and data providers in order to facilitate business operations.

Performed Allscripts Sunrise 5.5 ClinDoc training to RNs, NA/PCT's, PAs/NPs and MDs. Developed instructional materials. Provided Activation support.

Developing web and desktop applications in Python and PHP, using MySQL for database interfacing. Built and maintained multiple websites using a variety of frameworks ranging from Ruby on Rails to ASP.NET. Provided Microsoft Windows Server Administration and Network Support. Supported a variety of ERM software including Fishbowl Inventory, Quickbooks, and Aldelo POS.