Machine Learning Scientist

Sintela is a deep tech company that specializes in Distributed Acoustic Sensing (DAS). DAS is a technology that can transform an ordinary hundred-mile-long optical fiber into a hundred thousand acoustic sensors—equivalent to microphones, accelerometers, and geophones—that are distributed every few feet along the entire fiber length. DAS enables real-time, large-scale monitoring of critical infrastructure, the natural environment, and the energy and transportation sectors to name a few.

Sintela is the global leader in DAS, serving clients across the world. Applications include border security, perimeter intrusion detection, and monitoring of oil and gas pipelines, mining, roads, and rail. Among our most significant applications is the use of DAS to detect illegal activity along US borders.

Central to Sintela’s success is the Autonomous Detection Group, which develops and maintains machine learning and signal processing algorithms for DAS. Sintela is recruiting a

Machine Learning Scientist

to join the Autonomous Detection Group.

What You will Do

As a Machine Learning Scientist, your role will be to explore state-of-the-art and bourgeoning machine learning methods for DAS data. Specifically, you will:

• Comparatively and quantitatively evaluate various deep learning / machine learning training paradigms, such as supervised, unsupervised, semi-supervised, self-supervised, reinforcement, contrastive, and physics-informed learning.
• Comparatively and quantitatively evaluate deep learning / machine learning model architectures—including convolutional, recurrent, and transformer-based neural networks, diffusion models, and autoencoders—across tasks such as classification, generation, and latent space representation learning.
• Comparatively and quantitatively evaluate various methods of transfer learning, parameter-efficient fine-tuning (LoRA), knowledge distillation (teacher-student), domain adaptation, zero-shot, few-shot, and N-shot learning.
• Target and quantitatively evaluate the implementation of models in real-time on the edge.
• Work in various data domains—space, time, frequency, space-time, frequency-time, and frequency-space.
• Explore and adapt foundation models across various data domains.

Minimum Requirements

• Have an existing or can obtain and maintain a security clearance with the Department of Homeland Security
• Graduate degree in computer science, electrical engineering, physics, or similar and 2 years of experience.
• A strong foundation in – and passion for – machine learning.
• A track record of publishing in academic journals and conferences, such as ASA, ASG, SEG, Optica, IEEE, SPIE, CVPR, and ICML.
• Strong understanding of machine learning evaluation methodology, including performance metrics (ROC, PR curves, F1, confusion matrices) and experimental design for imbalanced classification problems.
• Strong foundation in machine learning related mathematics, principles, and theories.
• Statistics / probabilistic modeling — detection theory, probability of detection vs. false alarm, Bayesian reasoning.
• Experience with deep learning frameworks — PyTorch, TensorFlow
• Experience with MLOps tools (e.g., MLflow, Docker, Kubeflow, Airflow, Kubernetes).
• Excellent programming skills with Python and associated ML libraries.
• Experience with software version control tools such as Gitlab.
• Technical documentation experience.
• Will work well individually and in collaboration with an international (primarily US-UK) team.
• Demonstrate integrity as well as physical and cyber security consciousness.
• Experience with Linux systems.

Other Competencies of Interest

Knowledge and skills in the following domain areas are additionally of interest:

• Go programming language.
• Cuda programming.
• PostgreSQL.
• Data management / data science.
• Signal processing.
• Distributed Acoustic Sensing.
• Experience with cloud platforms (AWS, GCP, Azure).
• Digital signal processing (filtering, FFT, spectral analysis)
• Data fusion.
• Physics and mathematics.
• Seismology.
• Conventional image processing (e.g. shape detection).

Benefits

• Enjoy working as part of an international (primarily US-UK), multi-disciplinary team of scientists/engineers in a friendly, informal and fast-paced development environment delivering robust Autonomous Signature Classification workflows.
• Hone your expert skills and experience the satisfaction of pitting them against a range of temporally, spatially and spectrally diverse signatures.
• Witness the product of your efforts transition rapidly from concept to operational deployment and delivery of real-world effect, often thereby directly contributing to the prevention of illegal activity.

Employment Conditions

Tight collaboration and the sensitivity of some signature datasets demand an on-site working policy.

Candidates must be willing to undergo the Homeland Security Full Background Investigation. Employment is contingent on satisfying this security check.