Director of Engineering-- Data Infrastructure

About the Role

DreamVu builds data infrastructure for Physical AI — capturing, processing, and annotating real-world multimodal data used to train humanoid robots and embodied AI systems.

We're looking for a Director of Engineering to own the infrastructure that powers our data operations. This is a pure engineering leadership role focused on reliability, scalability, throughput, and cost efficiency. The ideal candidate has run large-scale cloud or data infrastructure in a demanding production environment — and managed the cost of doing so — and wants to apply that experience to one of the most consequential data problems in AI.

What You Will Own

The role spans three areas:

Pipeline operations — end-to-end ownership of the data pipeline, from raw sensor capture through pre-processing, automated pre-annotation, and human-in-the-loop annotation and QA. Full pipeline details are shared during the interview process.

Cloud infrastructure and cost — ownership of the cloud platform the pipeline runs on: provisioning, scaling, reliability, and the cost efficiency of it all. You'll treat spend as a first-class engineering metric — right-sizing compute and storage, controlling unit economics as data volume grows, and keeping cost-per-hour-of-data on a downward trend.

Systems integration and tooling — hardening and stabilizing new pipeline components into production, plus the monitoring, alerting, QA frameworks, dataset management, workflow tooling, and operational documentation that support consistent delivery.

Responsibilities

• Own pipeline reliability, throughput, and quality SLAs end-to-end
• Own cloud infrastructure spend — forecast, monitor, and continuously optimize cost as data volume scales
• Build and lead the engineering and operations team as the pipeline grows
• Define and manage how new pipeline components become stable production systems
• Maintain a clear view of technical debt and a plan to address it
• Establish engineering practices — code standards, review processes, incident response, documentation
• Translate data capture roadmaps into capacity plans and delivery schedules
• Maintain visibility into pipeline health and cost for the team and leadership
• Coordinate with capture operations ahead of new data collection campaigns

What We're Looking For

• 8 years in engineering, including 3 years in a leadership role — ideally in cloud services, data infrastructure, or large-scale platform engineering
• Track record running high-throughput, high-availability production systems on AWS, GCP, or Azure
• Demonstrated ownership of cloud cost — you've managed infrastructure budgets and driven down unit economics at scale
• Strong operational instincts — you define SLAs, instrument systems proactively, and treat instability as an engineering problem to be solved
• Solid Python fundamentals and comfort extending an existing codebase
• Experience building and managing small, high-ownership engineering teams under delivery pressure
• Strong planning and communication skills — comfortable reporting status and cost clearly to the CTO
• Nice to have: experience with video or multimodal data at scale (depth, point clouds, sensor streams)
• Nice to have: experience with annotation platforms or human-in-the-loop tooling

What Makes This Role Distinctive

DreamVu's pipeline processes multimodal sensor streams at scale across a complex, multi-stage workflow. The challenges — throughput, reliability, cost efficiency, and quality at volume — are the same class of problems found in demanding cloud data platforms, but applied to a domain at the frontier of AI. For an infrastructure leader who wants high ownership and meaningful scope, this is an unusually substantive opportunity.

Location & Structure

Based at our Philadelphia office, with regular coordination with our Hyderabad team. Reports to the CTO.