Machine Learning Researcher

Key Skills

Deep Learning, Graph Learning, Machine learning techniques, Natural Language Processing (NLP), Neural Networks, NumPy, PyTorch, Reinforcement Learning, Scikit-learn, SciPy, TensorFlow

Job Description

About the Team

STR’s Analytics division researches and develops advanced analytics and machine learning-based solutions to solve challenging problems related to national security. Our team consists of passionate and motivated engineers with advanced degrees in engineering, computer science, mathematics, and data science, who are seeking opportunities to use their deep technical knowledge and creativity to tackle some of the hardest problems that our customers face. Our projects span multiple different data modalities and incorporate advanced algorithms, deep learning, and statistical techniques to uncover patterns in social media, structured and unstructured text, time series, geospatial, and imagery data, and must operate under challenging constraints not typically found in the commercial world. The tools and technologies we develop have real world impact and US Government analysts use them to extract and enrich intelligence information around the globe.

The Role:

As a Machine Learning Researcher, you will utilize State of the Art (SOTA) deep learning methods to work with disparate and unlabeled multimodal data sources to create a knowledge engine that can be easily queried by humans. You will collaborate with a diverse team of people that includes, software engineers, algorithm developers, and academics. Your work will leverage cutting-edge learning techniques and demonstrably improve human-machine collaborative analysis for a real-life scenario.

What you will do:

• Work with unlabeled multimodal datasets to extract concept representations and their relationships.
• Implement methods for learning object, property, affordance, and activity representations in a shared latent space for modality-spanning concepts.
• Utilize the joint latent space representations to resolve inconsistencies within datasets and across modalities.
• Correlate similar/dissimilar concepts across modalities.
• Implement a continual learning approach that utilizes armed bandit techniques from RL to cast model retraining under data drift as a balance between exploration and exploitation.
• Collaborate with a diverse team that includes academics and industry partners.

Who you are:

• Hold a M.S. or Ph.D. in Machine Learning or Computer Science with concentration in Machine Learning (deep learning/reinforcement learning).
• U.S. citizen
• Experienced in using SOTA machine learning to solve real world problems, especially in settings where rich multimodal data is involved.
• Experienced with deep learning and optimization techniques in joint latent spaces.
• Experienced with scientific computing packages in python such as NumPy, SciPy, scikit-learn and tensor library PyTorch.
• Familiar with unsupervised and semi-supervised learning techniques.
• Familiar with version control (Git) and collaboration tools (JIRA, Confluence, or related).
• Capable of obtaining and maintaining a security clearance.

Nice to have:

• Deep learning experience that spans the domains of Computer Vision, Natural Language Processing (NLP), and to a more limited extent Reinforcement Learning (RL).
• Experienced with Graph Neural Networks and graph-based analyses (e.g., relation extraction, knowledge graph creation, hierarchical ontology construction, etc).

STR is a growing technology company with locations near Boston, MA, Arlington, VA, near Dayton, OH, Melbourne, FL, and Carlsbad, CA. We specialize in advanced research and development for defense, intelligence, and national security in: cyber; next generation sensors, radar, sonar, communications, and electronic warfare; and artificial intelligence algorithms and analytics to make sense of the complexity that is exploding around us.