Staff MLOps Engineer

Description

About AiDASH

AiDASH is leading the PreventionFirstmovement for electric utilities and transforming grid resilience through its pioneering platform that unifies vegetation, asset, storm, and wildfire intelligence. Powered by SatelliteFirst Inspection & Monitoring, AiDASH delivers comprehensive visibility across the entire grid at the right frequency and budget, using the right data modality. More than 200 customers trust AiDASH to keep the lights on, spend where it counts, and defend every decision, Securing Tomorrow across every mile of the grid. Learn more at www.aidash.com.

The PreventionFirst movement is growing, and so is the recognition behind it. In 2026, Forbes named AiDASH one of America's Best Startup Employers for the 4th consecutive year, and TIME included AiDASH among America's Top GreenTech Companies for the 3rd year in a row. Deloitte Technology Fast 500 ranked AiDASH No. 12 in the San Francisco Bay Area, and No. 59 overall in their selection of the top 500 for 2024.

Join us in Securing Tomorrow Together!

Responsibilities

• Our satellite and AI-powered products are changing how utilities see and protect the grid, and we're gearing up for aggressive customer growth over the next two years. That growth runs on AI at scale. We're building a brand-new Data Inference AI Pipeline team to deliver it, and we're looking for a Staff MLOps Engineer to architect its operational backbone from the very first line of infrastructure code.
• As the very first engineer on this team , you'll start with a blank canvas and a rare chance to build something that lasts. You'll shape how our models are deployed, how the pipeline scales and recovers under load, and what it costs to run. Just as importantly, you'll help define the culture, engineering standards, and ways of working that every future teammate inherits. This is a hands-on, high-leverage individual contributor role: your decisions will set the foundation the entire team builds on, and you'll be the go-to technical authority on MLOps as the team grows. You'll report to the Director of Engineering for the Data Inference AI Pipeline team.
• Our pending acquisition by Schneider Electric will accelerate our mission as we bring our combined offering to our joint customer base, and the infrastructure you build will power outcomes at an even greater scale.
• Location: This is a hybrid role based in Palo Alto, CA, requiring 2 days per week in the office.
• How you'll make an impact:
• Deployment
• Own the model deployment pipeline end-to-end: packaging, versioning, rollout, and rollback for ML models moving from training to production
• Build and maintain deployment infrastructure on SageMaker (endpoints, batch transform, multi-model/multi-container hosting), and evaluate where alternatives such as self-hosted serving or other managed options are a better fit
• Set the CI/CD pattern for model releases, including automated testing, staged rollout, and canary/shadow deployments
• Scalability & elasticity
• Design inference infrastructure that scales seamlessly with real-time and batch demand
• Right-size compute (CPU/GPU/Inferentia or equivalent) for each model and workload to balance latency and cost
• Load-test the pipeline and build capacity plans that stay ahead of customer growth milestones
• Cost
• Own the cost-per-inference metric: instrument it, report on it, and drive it down as volume scales
• Build cost-awareness into architecture decisions, including autoscaling policies, spot/on-demand mix, and model size/quantization trade-offs
• Failover & resilience
• Design for graceful degradation and failover across regions and availability zones for inference serving
• Define and test disaster-recovery procedures, and run regular failure-injection and chaos exercises
• Build monitoring, alerting, and on-call runbooks for model-serving infrastructure, covering drift, latency SLOs, error rates, and infra health
• Cross-team leadership
• Partner with Data Science to bring new models to production on a platform built to support them
• Partner with the Web Application team where inference results power customer-facing features, keeping latency and reliability expectations aligned
• Serve as the technical authority on MLOps practices as the Data Inference AI Pipeline team grows
• What success looks like in your first 6 months:
• Model deployment is a repeatable, automated process across models
• Inference infrastructure has proven it scales up and down smoothly under real load
• Cost-per-inference is measured, visible, and trending down as volume grows
• A tested failover and disaster-recovery plan is in place for the inference pipeline, with at least one failure-injection exercise completed
• Data Science can ship new models into production quickly and independently

Requirements

• Minimum Qualifications
• 7-10+ years in software/ML engineering, including 3+ years focused specifically on MLOps or ML infrastructure
• Hands-on production experience with AWS SageMaker (or an equivalent managed ML platform) for model deployment and serving
• A proven track record designing systems for elasticity/autoscaling, cost optimization, and failover/resilience at production scale
• Strong expertise in containerization and orchestration (Docker, Kubernetes) and in CI/CD built for ML workloads, including model versioning and automated testing and rollout
• Experience instrumenting and owning cost and performance metrics for the infrastructure you're responsible for
• The ability to set technical direction as a Staff-level IC, building buy-in across Data Science and other engineering teams
• High ownership mindset, low ego, and a collaborative spirit, with a passion for building platforms that help others ship faster
• Preferred Qualifications
• Experience with large imagery, video, or 3D sensor and point-cloud data pipelines or other high-volume unstructured data at scale
• Experience with multimodal data (imagery, sensor/time-series, geospatial) beyond purely textual or tabular data
• Experience building MLOps practices from scratch on a new team
• Familiarity with model monitoring and drift-detection tooling, and GPU cost-optimization techniques (quantization, batching, Inferentia/Trainium or equivalent)
• A background in a regulated or enterprise B2B domain such as utilities, energy, infrastructure, healthcare, or finance
• Relevant AWS/Kubernetes certifications or open-source MLOps tooling contributions
• What you'll love:
• Comprehensive Medical, Dental, and Vision Coverage: 100% coverage for employees and 80% for their spouses and children
• Health Reimbursement Account (HRA): 100% funded by AiDASH to cover medical deductibles
• 401(k) Plan: Begin contributing after two months of employment to prepare for your future. Currently, no company match is offered
• Parental Leave: Supportive parental leave with 16 weeks for primary caregivers and 4 weeks for secondary caregivers
• Generous Vacation Policy: Accrue 20 vacation days per year, plus enjoy an additional flex holiday to celebrate whatever feels most important to you
• Winter Break: From December 25th through January 1st, we give everyone time off to recharge and enjoy time with family and friends!
• We are proud to be an equal-opportunity employer. We are committed to embracing diversity and inclusion in our hiring practices, and we promote a work environment where everyone, from any race, color, religion, sex, sexual orientation, gender identity, or national origin, can do their best work.
• We offer a competitive annual pay range of $230,000 to $270,000 for this full-time position, which includes a base salary and bonus based on performance. This range reflects the anticipated annual pay range (base + bonus) for new hires. We strive to ensure our compensation packages are equitable and aligned with industry standards. Your recruiter can share more about compensation during the hiring process.
• We are committed to providing an inclusive and accessible interview experience for all candidates. Please let us know if you require any accommodation during the interview process, and we will make every effort to meet your needs.