Peraton Labs delivers innovative solutions and revolutionary new capabilities to solve the most difficult and complex challenges for government agencies, utilities, and commercial customers. With a distinguished heritage tracing back to Bell Labs, Bellcore, and Telcordia, our experts pave the way. Peraton Lab’s cybersecurity research protects mission-critical systems and national cyber infrastructure through a broad range of initiatives in computer network defense, secure-by-design techniques and cyber operations and experimentation platforms.

The Machine Learning/Data Science Research Engineer will participate in an end-to-end design and implementation of deep learning solutions for application in network/graph representation learning (NRL) and graph construction/generation/completion.  

The ideal candidate will be able to start from the end users' needs and understand how they drive performance and infrastructure requirements. The ideal candidate has experience converting customers' needs into user interface requirements and baselining and scaling ML solutions using CI/CD best practices. 

Roles and Responsibilities: 

• Frame the problem, research current state-of-the-art ML solutions, and assess their suitability to the problem at hand.  
• Leverage deep learning frameworks (e.g., PyTorch, Tensorflow) to develop deep learning solutions, tailoring them to the application area.  
• Develop extract-transform-load (ETL) pipelines to curate large-scale ML datasets.  
• Leverage graph neural networks (GNN) to construct embedded/latent representations for downstream tasks. 
• Train ML models using automated hyper-parameter tuning frameworks. 
• Design data model to capture real-world phenomena using graph data structures in a space/computationally efficient design

Required Skills:

• Minimum of 8 years with BS/BA; Minimum of 6 years with MS/MA; Minimum of 3 years with PhD
• Minimum of 2 years team lead experience
• Experience training conventional models from scratch (e.g., ordinary least squares (OLM), random forests, support vector machines (SVM), boosting methods, etc.) 
• Personally trained a deep neural network (DNN): either trained a DNN from scratch OR leveraged transfer learning techniques to further tune a pretrained DNN to a specific target domain. 
• Developed data wrangling/ETL transforms using python Pandas package. 
• Experience with Git version control. 
• US Citizenship
• Top Secret clearance with the ability to obtain SCI

Desired Skills: 

• Experience training a deep neural network from scratch 
• Experience finding white-papers applicable to problem area and tailoring provided code to application domains 
• Experience with Generative Modeling (e.g., GANs) 
• Experience with Reinforcement Learning (RL) solutions 
• Network engineering experience (e.g., knowledge of OSI/TCP/IP models, network infrastructure) 
• Experience developing on GNU/Linux-based operating system 
• Experience with MLOps tooling/workflows 
• Experience using containers (e.g., docker) 
• Experience with orchestration solutions, including Docker Swarm, Kubernetes, Terraform