Senior Machine Learning Engineer - Active Learning

Sense•Cambridge, MA

22h•Hybrid

About The Position

Sense is a climate tech company that uses electric signals in homes and on the grid to provide real-time insight into energy use, devices, and faults. Its load disaggregation and fault detection models run on meters and support applications including helping people reduce electricity use and helping utilities prevent wildfires. The Senior Machine Learning Engineer will own the loop from data to labels to improved models, using active learning and weak supervision to generate labels, determine what data should be labeled next, and turn field errors into new labels. This is a hands-on role involving coding, model training, data analysis, and experiment documentation. The engineer will work with data scientists, annotators, product, and engineering teams, and direct annotation-tool development and annotator work. Success is measured by reducing time and cost per annotation while maintaining label quality at or above targets.

Requirements

MS or PhD in Computer Science, Electrical Engineering, Physics, Statistics, or a related field, or equivalent experience.
At least 5 years of experience in data science or machine learning, including at least 2 years building or managing labeled datasets for machine learning.
Hands-on experience applying active learning to real data, including selecting a sampling strategy and measuring its impact.
Experience finding rare or difficult examples in large datasets using model errors, uncertainty, or similarity search.
Experience training deep learning models and using weak supervision or model-assisted labeling.
Experience measuring label quality and correcting label errors.
Experience with Python, SQL, and data pipelines.
A track record of working with other teams, such as product, software, or hardware.
Able to work from the Cambridge, Massachusetts office at least 2 days per week.
Work authorization required.

Nice To Haves

Experience with time series, audio, sensor, or other signal data.
Experience leading an annotation team or directing development of annotation tools.
Experience using embeddings from large or foundation models for data analysis.

Responsibilities

Design and run active learning to select which homes, devices, and time periods to label based on model uncertainty or errors.
Turn bad detections, field accuracy results, and user corrections into ground truth, and track whether retrained models improve.
Find rare events and difficult cases in home data, including uncommon or mislabeled devices, and add them to labeling and evaluation sets.
Own the ground truth plan, including which data to label, how much, and in what order.
Direct annotators by setting priorities, tracking throughput, and providing guidance.
Set data quality standards.
Direct development of the annotation tool and add features that enable collection of new and better data, including model-assisted features.
Measure and improve label quality using metrics, outlier detection, human review, and other methods.
Work with data engineers and ML operations to version labels and track which labels trained which model.
Keep evaluation sets separate from training data so offline results predict field performance.
Review code, provide peer feedback on research and experiments, and contribute to shared documentation.
Help mentor and coach less experienced team members.