The posting
About Ironsite
Ironsite is building the intelligence layer for the physical world. We design our own wearable hardware, deploy it alongside craft workers, and transform a shift's footage into a next-morning report. Our internal team and purpose-built models label the data overnight and deliver actionable insights to superintendents by 5 AM.
We are accelerating the speed, efficiency, and predictability of construction, especially for complex, mission-critical infrastructure projects, including data centers, LNG facilities, sports stadiums, hospitals, and other large-scale developments, by training AI models on egocentric construction footage and labor productivity data. We are built with a pro-worker philosophy at our core: we believe technology should empower the workforce, not replace it.
Ironsite is deployed across several of the largest active construction projects in the country. To date, we've captured more than 100,000 hours of construction footage across seven states, now process thousands of hours of site activity every day, and maintain a worker opt-out rate below two percent. This is enabled by a workforce-first architecture that anonymizes devices, captures no audio, and never releases raw video.
Ironsite is backed by leading investors (8VC, South Park Commons, Saga Ventures) and prominent operators across technology and construction, including Eric Schmidt, Jeff Dean, Jeff Rothschild, Mark Leslie, Scott Wu, Eric Glyman, Karim Atiyeh, Russell Kaplan, and others, alongside over a dozen construction industry operators who have joined us as partners in building this.
Longer term, we believe Ironsite is the foundation for what construction becomes in the next decade. We see a future where craft workers are more valued, more visible, and better paid for the skill they bring, and where the industry finally has the intelligence layer that makes autonomous construction possible.
The Role
We're hiring a Head of ML Engineering to own how our research team's models get tested, labeled, shipped, and run in production.
Research builds the models. You'll build everything around them: the systems research uses to A/B test models on live jobsites, the tools our labelers use every night, the data tooling research trains from, and the orchestrator and inference stack that deliver reports by 5 AM. You'll start as a hands-on tech lead and grow the team from there.
What You'll Own
Experimentation for Research. Build what research needs to A/B test models in production: run two models side by side on real sites, compare their output, and decide what ships. When a test wins, roll it out everywhere. Labeling. Own Tagging Interfaces, the tool our labeling team uses every night. Pre-label footage with models so labelers correct rather than start from scratch, decide what gets labeled next based on what research needs, and track throughput and label quality. Data Tooling for Research. Make 100,000+ hours of footage easy to use: dataset versioning, search across footage and labels, and fast assembly of training sets. Product Features. Ship ML-backed features in the superintendent report. When a customer gets a bad report, trace it to the model, the labels, or the pipeline and fix it. Get models working on each new jobsite during onboarding. MLOps, Orchestration & Inference. Own the orchestrator that runs the nightly pipeline, plus model deployment, versioning, rollback, and inference. Watch output quality by site and catch drops before customers do. Own GPU and cloud spend, and absorb footage and device changes from the hardware team without breaking the pipeline.
Challenges You'll Solve
Fast experiments on live reports. Research wants answers in days, not weeks. Customers need every 5 AM report to be right. You'll build a system that gives research both. Labeling as the bottleneck. Every model improvement starts with labeled footage. You'll get more useful labels out of every labeler hour, month over month. More sites, same deadline. Each new jobsite adds footage, but the report is still due at 5 AM. You'll scale the pipeline without cost growing at the same rate. Doing and building at once. You'll write much of the code yourself while hiring the people who take it over.
What We're Looking For
7+ years of software engineering, with at least 3 years building production ML systems. Experience leading a small team as a tech lead or manager while still shipping a lot of code yourself. You've built experimentation or A/B testing systems in production. You've built internal tools that non-engineers use every day, such as labeling tools or operations tools. Hands-on experience with MLOps and GPU inference. You can debug across the stack: from a wrong number in a customer report back to the model, the data, or the pipeline. Clear, concise communication with researchers, labelers, and customers.
Preferred Qualifications
Experience with video data at scale. Experience with labeling tooling: pre-labeling, active learning, or QA workflows. Frontend or product experience, since Tagging Interfaces is a product with daily users. Experience working alongside a research team at an ML startup. Experience with data from device fleets or other hardware. Familiarity with construction or other physical-world operations.
Location, Compensation, & Perks
San Francisco Bay Area (on-site) Base salary: $200k-$330k per year, commensurate with experience Meaningful founding-team equity. Full benefits including health, dental, vision, and 401(k) with 6% match Daily catered lunch Office in San Francisco, next to Oracle Park and the Caltrain
Final compensation is determined by experience, location, and level.
