Natural Language Processing (NLP) Engineer medCPU, Inc. - Pittsburgh, PA medCPU, a software company specializing in Clinical Decision Support (CDS) software for the healthcare industry, is looking for an NLP Engineer to join our growing Pittsburgh team! The NLP engineer will design, implement, test, and maintain medCPUs clinical information extraction systems. Works with an interdisciplinary team of clinical professionals, software engineers, and decision support analysts to deliver value on key healthcare outcomes such as quality, safety, and cost-effectiveness of health care delivery. Fast-paced, start-up environment, with immediate impact on patient care. At medCPU, we work in small teams (4-6 people) focusing on an aspect of the overall project. Our engineers, QA and product analysts work together to build solutions that meet the needs of our customers and internal operations teams. Our engineers evaluate new approaches, frameworks, and libraries on a regular basis and strive to ensure we're using the best tool for the team and the job at hand. Responsibilities
* Design, develop, test, and maintain NLP pipeline for real-time clinical information extraction. * Develop and evaluate methods for real-time classification and inference from complex streams of NLP annotations. * Develop and evaluate clinical NLP components for new use cases, within a standard pipeline-based framework (e.g. tokenization, parsing, POS tagging, entity recognition, relation extraction, etc). * Work closely with clinical terminologist to ensure adequate coverage, structure and detail of medCPU terminologies for NLP components. * Work closely with decision support specialists and engineers to ensure adequate performance (both speed and accuracy) of NLP systems to support medCPU decision support goals. * Participation in the full software development cycle using Agile methodology Knowledge, Skills, and Attributes * BS, MS, or PhD degree in Computer Science or related field * 3+ years of experience with developing NLP or text-mining applications * Knowledge of and experience with one or more NLP pipeline architectures (UIMA, GATE) * Knowledge of clinical NLP pipelines (e.g. Apache cTAKES), standard terminologies (e.g. SNOMED), and data standards (e.g. HL7, FHIR) * Ability to evaluate NLP systems using standard performance metrics, and to iteratively increase performance using error analysis * Previous experience developing statistical and machine learning language models a plus * Ability to write well documented and clean code using software engineering best practices Additionally, the following skills would be desirable: * Java 8 and Spring Framework * Event sourcing/CQRS * RESTful API development * RDBMS such as MSSQL, Postgres, etc. * Unit and integration testing * Git-based workflows * Cloud technologies and architecture such as AWS Our Location in Pittsburgh's Southside This position is full-time, based in our office at 2100 Wharton Street, Pittsburgh, PA on Pittsburgh's 'South Side'. The South Side is a vibrant, youthful community and a weekend hotspot with a mix of tech startups, hip eateries, eclectic shopping and watering holes. Pittsburgh is a beautiful, affordable city, and frequently among the top five most livable cities in the USA. Increasingly known for its burgeoning technology sector, Pittsburgh is attracting top tech talent from around the world. Additional Information At medCPU, we enjoy a casual work environment, with a 14-foot shuffleboard table providing a friendly competitive outlet for those times when you just need a break. The building also provides an on-site café and gym. * Benefits: Includes full benefits (Medical, dental, vision, retirement, paid time off, and more). * Visa sponsorship: Not available for this position. Please contact: Nick Dudukovich Recruiter, Talent Acquisition UPMC Enterprises [email protected]<mailto:[email protected]> Rebecca Jacobson, MD, MS, FACMI VP Anaytics, UPMC Enterprises Bakery Square 6425 Penn Avenue #200 Pittsburgh, PA 15206 Office: (412) 863-4946 Cell: (412) 606-7105 [email protected]<mailto:[email protected]>
