Data Scientist

San Leandro · Headquarters · USD 100,000-160,000 · 1-3 years experience · Onsite · Posted Oct 1, 2026

At a glance

SirenOpt helps manufacturers make better, safer, and more reliable micro- and nano-materials. These materials are the building blocks of critical sectors of the global economy such as batteries, computer chips, aircraft components, and power systems. But, surging material demand and growing complexity are pushing production to unprecedented scales and speeds, leaving manufacturers effectively flying blind. Small, undetected variations during production lead to wastage, lower performance, higher costs, and safety risks.
SirenOpt is changing this by developing a manufacturing intelligence platform that non-destructively probes materials during production, revealing critical internal information without damaging them. Using a novel combination of cold atmospheric plasma, physics-informed machine learning, and predictive analytics, SirenOpt generates unique, real-time material fingerprints that capture material signals not accessible through conventional measurement techniques. These insights give manufacturers unprecedented visibility into how materials behave as they are made.
We turn hidden data into actionable intelligence to help manufacturers reduce variability and thus increase yield and performance. The technology can be deployed as a standalone tool or integrated directly into factory production lines. SirenOpt is currently deploying early versions of its platform with some of the largest industrial manufacturers in the world across North America, Europe, and Asia.

SirenOpt is seeking a Data Scientist for their Applications Engineering team to build and deploy machine learning models that turn complex sensor signals into actionable predictions about material properties.
This forward-deployed, customer-adjacent role involves executing proof-of-concept studies with customer samples, validating model performance on novel materials, and collaborating with engineering teams to move models into production.

Build, calibrate, and validate predictive models that map sensor signal features to material properties
Design and evaluate new model architectures and featurization strategies suited to small-data, high-dimensional scientific datasets
Apply methods including regression, dimensionality reduction, probabilistic modeling, anomaly detection, and physics-informed ML
Develop testing and validation frameworks for model performance, including uncertainty quantification and out-of-distribution detection
Characterize model robustness across sample types, process conditions, and instrument configurations
Prepare models and documentation for handoff to the software engineering team for production deployment
Analyze datasets from customer proof of concepts
Compile technical reports and supporting materials to deliver to customers
Translate findings and stakeholder feedback into model improvement roadmaps

B.S. in Data Science, Statistics, Applied Mathematics, or a related quantitative science field with 3-5 years of applied ML/data science experience; or M.S. with 1-3 years
Hands-on experience building and validating predictive models (supervised and self-supervised) in Python
Ability to analyze multivariate, high-dimensional datasets and perform feature engineering and selection
Solid grasp of statistical modeling: uncertainty quantification, regularization, covariate analysis, and feature importance methods
Strong communicator; comfortable presenting technical findings to both technical and non-technical audiences

Ph.D. degree
Experience working with time-series, spectroscopic, or other sensor-based signal data
Prior work in manufacturing, materials science, energy storage, semiconductors, or another physical science domain
Prior customer-facing or applications engineering experience in a technical product company
Experience deploying models in production software environments
Familiarity with data pipeline development (PostgreSQL or similar)
Fluency in Mandarin Chinese, Japanese, German, Korean, or another key stakeholder language