Machine Learning Engineer

About Us

Sentinel Devices is at the forefront of decentralized AI / ML for industrial equipment and process data. We build data collection and monitoring platforms for industrial equipment that can operate completely off of the cloud - all data storage, processing and decision-making happens entirely locally.

This means our devices are used to monitor hard-to-reach or security-sensitive assets - we make reaching these assets cost-effective for our customers.

About the Role

We develop automated pipelines enabling our devices to build artificial intelligence / machine learning (AI / ML) models from time series data using minimal computational resources.

For our systems, data sources are typically sensor and control values collected from assets in the field, such as valve positions, temperatures, pressure values, etc.

Our AI / ML pipelines need to be able to train themselves from scratch using data specifically marked by the user, and need to be able to adapt to user feedback and re-train themselves on the fly.

Once built, these models need to be deployable in a highly automated fashion for monitoring, identifying, and triangulating anomalies in the source data.

Most importantly, these pipelines need to function with no supervision by us - once a system is deployed, we have to assume we'll never touch it again.

As a Lead AI / ML Engineer, you will :

• Work closely with external partners, stakeholders and customers to understand their operational and data needs and translate these into requirements for our core AI / ML systems.
• Spearhead the design, development and implementation of the core AI / ML systems in our platform responsible for identifying anomalies in industrial time series data, and backtracking these anomalies to identify potential root-cause signals and events.
• Coordinate with other company employees and contractors to deliver the envisioned core AI / ML systems, developing roadmaps, implementation plans and strategies to guide and coordinate company personnel.
• Develop and engineer the core AI / ML systems to deal with arbitrary time series data streams from a variety of sources and system types.
• Use data provided by our customers or partners to test, refine and debug the core AI / ML systems.
• Use data provided by our customers or partners to identify edge cases and failure modes in the core AI / ML systems and either mitigating them or putting in guardrails to ensure they are not an issue in production deployments.
• Be forward-thinking and strategic in your algorithm choices as well as your system design, as you'll have to design these AI / ML systems to operate, potentially for years, with minimal human intervention.

This is a full-time on-site position, which means that you will be required to work within the company's office space in Atlanta, GA.

Because of specific contract requirements with government sponsors, this role requires U.S. citizenship.

Technologies in our stack :

• Python (a LOT of Python)
• C / C++
• RESTful APIs
• JWT authentication
• MQTT
• Docker

The following skills & experience are required :

• Experienced and comfortable working in Linux
• Experience programming in Python or C / C++
• Experience in developing and applying machine learning to practical engineering problems
• Experience with code versioning systems (git)

The following skills are not required, but would be a good plus :

• Previous experience in industrial outlier detection or anomaly detection
• Experience with industrial time series data (sensor data, machine data, etc.)
• Experience in dealing with real-world industrial data (missing data, corrupted data, seasonal data, etc.)
• Experience developing AI / ML systems using limited training data
• Experience in embedded intelligence applications (AI on a Raspberry Pi, etc.)
• Experience with containerization & working with Docker (experience with Balena strong plus)
• Experience working or reverse-engineering IoT protocols
30+ days ago