Medford, Massachusetts, United States
James Clarke is an internationally recognized expert in natural language processing (NLP). Experienced in translating ideas and techniques from the research literature into working implementations and building them further. Develops methodologies and frameworks to aid building high-level text understanding applications. This includes frameworks to integrate and scale natural language processing and machine learning components. His research has focused on statistical natural language processing and machine learning. He has worked on a variety of problems including: information extraction, extracting semantic information and actionable knowledge from text, automatic text summarization, and machine translation. In 2007 he received the best paper award at the Conference on Empirical Methods for Natural Language Processing (EMNLP) for his PhD work on incorporating discourse information into automatic text summarization via constraints. James is a full-stack programmer. He is capable of developing broad applications that include: novel text analytics and machine learning systems and building, web application front-ends, data processing workflows, data storage solutions, and software services. Programming languages of choice: Java, Python and Unix tools; HTML/CSS/Javascript with PHP or Python for front-end work. Technologies: Apache Thrift, MongoDB, Condor, Solr, Data Warehousing, Git, Subversion, Maven, web.py, RESTful APIs, Service-oriented architecture Specialties: Natural Language Processing, Computational Linguistics, Automatic Summarization, Machine Learning, Structured Learning, Information Extraction, Integer Linear Programming Model Building, Artificial Intelligence
Polytomic is a new way to sync data from your databases, data warehouses, and spreadsheets to business systems like Salesforce, Marketo, HubSpot, Airtable, and others. Setup takes five minutes. We are backed by top investors including Y Combinator, Caffeinated Capital, Elad Gil, Jack Altman (Lattice CEO), Jeff Queisser (Box co-founder), Parker Conrad (Rippling CEO), and others.
Basis Technology provides multi-lingual text analytics capabilities and solutions. I managed an Agile/Scrum development team of software and research engineers prototyping and building new products and capabilities. As product owner, technical owner, and part of the development team, my responsibilities included: • Deciding the most valuable work to do sprint-by-sprint to achieve the broader company goals. • Designing how the work is done to fulfill our engineering architectural vision and research agenda. • Collaborating with product management to inform company strategy with respect to current and future NLP capabilities. • Hands-on development of specific functionality. • Mentor, develop, manage, and recruit top quality text analytics software engineers and researchers. • Act as a technical expert and advise other teams on NLP/ML topics and software architecture as needed. My team was responsible for the design and development of the Rosette Entity Resolver (RES) product, an entity-linking and cross-document coreference system. We developed RES from prototype and research phase through to 1.0 production release, and continue to support it.
Worked on a wide range of text analytics problems. Responsible for delivering technical solutions. Part of the leadership team shaping company strategy and product development. Balances research, engineering and business priorities. Architect of the research and development process and developed appropriate toolchain and NLP system. Developed prototype applications to showcase our technology's capabilities and data extracted to investors and clients.
Focused on natural language processing and machine learning. In particular methods for automatically interpreting language in the context of a world model; and exploring models for performing task-specific relation and information extraction, and database population. This research was applied to teaching computers to play card games by reading textual instructions written to accommodate human learning. Involved in multiple initiatives within the group to ease collaboration and access to resources. Exemplified by the creation the Curator architecture and system which provides a common programmatic interface to perform, retrieve and store annotations of texts. Reduces the amount of duplicate data processing and removes the burden of data processing, storage and maintenance from the user (researcher).