Description

Are you ready to join Leidos all-star team? Excited to apply your knowledge and expertise to solve a broad range of technical challenges and have real world impact? Enjoy working with a diverse and talent team? If the answer is YES! then, the Leidos AI/ML accelerator is a right place for you.

We are looking for a Senior Natural Language Processing (NLP) Researcher Scientist with a background that reflects designing, implementing, and analyzing approaches that integrate, adapt, and extend novel capabilities to state-of-the-art technical challenges in research and development (R&D) programs (basic and applied) as well as in direct client work. Candidates with a background in computational linguistics, human language technology, machine translation, automated speech recognition, machine translation, or linguistic data analysis, and experience participating in leading multi-disciplinary teams of researchers and engineers are strongly preferred. Work is located in Arlington, VA (Remote is acceptable).

This position requires a strong general knowledge of natural language processing. We need you to be proactive and carry the desire to push forward Leidos NLP expertise. You will need to able to develop NLP solutions with rigorous evaluation and clearly communicate such solutions with both the technical community and business customers.

This is your day to day:
- Lead and work with data scientists and machine learning engineers to adapt and scale NLP solutions for a variety of application domains.
- Design innovative and generalized NLP machine learning solution to extract or infer critical information from unstructured text or multi-modality data. Those solutions include but not limit to tasks like create/customize text embedding, classification, Named Entity Recognition, event extraction, topic modeling, and burst detection.
- Work closely with customers to gain thorough understanding of their decision making process and develop explainable NLP solutions to support or augment decision making.
- Design evaluation strategies for measuring both model performance and real world impact.
- Effectively present NLP work to stakeholders within and outside of Leidos as well as to the scientific community.
- Keep up with latest development in NLP, understand tradeoffs between classic and modern NLP models, specifically pre-trained language models.
- Serve as NLP subject-matter-expert for Leidos internal and external projects (e.g., DARPA or IARPA programs) and mentor junior NLP data scientists and researchers
- Produce technical white papers and scientific publications in NLP.

This is you:
- Masters with 10 years in computational linguistics, human language technologies, machine translation, or another NLP field or a Bachelor's degree with 2 additional years of applicable experience.
- Must be able to obtain a Top Secret security clearance with a polygraph security clearance.
- Publications/demonstrating strong technical results in a specialized subfield of NLP.
- Proven expertise building and validating NLP models with deep learning as well as traditional NLP methods. Hold strong understanding of the real-world advantages and drawbacks of those methods.
- Strong understanding of Transformer architecture and its variants, familiar with continue pre-training and fine-tuning language model with huggingface framework
- Ability to handle ambiguity, priorities, and competing objectives. Experience solving highly technical problems.
- Experience in the following areas: text processing and construction of corpora; processing of large text collections with standard NLP tools (e.g., scikit-learn, Spacy) for parsing, entity extraction, POS tagging, topic discovery and classification (such as sentiment analysis), and natural language understanding; tuning hyper-parameters of existing NLP models for domain-specific data sets; computational manipulation and analysis of natural language documents using statistical models; experimenting with large corpora for developing and testing advanced NLP algorithms.
- Attention to potential data and model biases and develop solutions to mitigate biases
- Ability to program in Python or other scripting language, comfort with work in a Linux environment on prime or in AWS Cloud
- Familiarity with common NLP and ML toolkits such as Stanford CoreNLP, OpenNLP, NLTK, scikit-learn, SpaCy, and Tensorflow/PyTorch.
- Ability to write clean, understandable code that follows leading industry standards and practices and is well-documented, and to build easily reproducible models
- Outstanding communication skills. Ability to communicate at all levels, including explaining sophisticated technical solutions to non-experts
- Open listener, ability to be open to many diverse voices and perspectives
- Experience providing mentorship and guidance to team members.
- Knowledge of state-of-the-art methods coupled with the creativity and intelligence to advance beyond them. Track record of active learning and creative problem solving.
- Working knowledge of AWS, Google Cloud Platform, Spark, data warehousing.
- Experience authoring research papers and getting those papers published

You will wow us even more if you have these skills:
- Experience winning and/or working on competitive contract R&D projects for a customer such as DARPA.
- Prior experience with many of the following models in the context of NLP problems and data: Deep Learning (CNNs, RCNNs, LSTMs, Transformers), GANs, Autoencoders, Reinforcement Learning, Siamese Networks, Logistic Regression, Linear Regression, Support Vector Machines, Hidden Markov Models, Conditional Random Fields, Latent Dirichlet Allocation.
- Hands-on experience with any of the following: Docker containerization, Kubernetes, Machine Learning Model Monitoring, Dataiku, Domino DataLab.

AIML
LInC

#Remote