Head of Dataset & Quality

Build AI; San Francisco (CA); On-site; USD 180,000 - 260,000; Full time

Benefits offered by this job

Competitive pay

Medical, dental, and vision

Housing subsidy

Relocation support

Wellness benefits

Daily lunch

Unlimited compute budget

Codex and Claude credits

Travel

Job summary

Build AI in San Francisco is seeking a data-centric quant researcher to define dataset strategy for physical labor tasks, design taxonomy, and the value function for incremental data. You will shape how we collect, label, and evaluate data to maximize informational gain.

You will collaborate with engineers and researchers to scale collection across real-world environments, set quality standards, and contribute across disciplines in a hands-on, in-person team.

Qualifications

• Quant background in statistics, decision science, or similar.
• Experience with dataset design, collection strategy, or large-scale data programs.
• Ability to argue about scaling vs fidelity with numbers.
• Familiarity with in-the-wild collection, video, or pose data (nice-to-have).

Responsibilities

• Define the ideal dataset for solving physical labor and the objective function on current data.
• Design taxonomy, golden sets, acceptance criteria, and audit for measurable quality.
• Create value function for incremental data and determine next data points.
• Translate strategy into collection plans, standards, and vendor execution.
• Avoid unnecessary process that reduces throughput and information gain.
• Align dataset decisions with model capability metrics and scorecards.
• Collaborate with Evals Lead to ensure dataset decisions inform models.

Skills

Quantitative background

Dataset design

Decision science

Advocating scaling vs fidelity

Job description

About Build AI

Build AI is the data hyperscaler for Physical AI. We're vertically integrated across hardware, manufacturing, logistics, collection, and model training to scale the physical labor dataset orders of magnitude faster than anyone in the world.

Job Summary

This is not “make the data prettier.” You own whether the dataset is actually solving physical labor. The objective is to maximize the bandwidth of net-new learnable information into the dataset: what we should collect, how we score what we have, and whether the next hour of collection adds information or just volume.

You define the ideal dataset, the objective function on the current dataset, the taxonomy, and the value function for incremental data. Collection strategy comes from those, including the bet that scaling simple, high-bandwidth capture (real workers, real jobs, a camera) beats slower high-fidelity setups.

A quant background is the default profile.

Key Responsibilities

• Define the ideal dataset for solving physical labor (coverage, diversity, what “done” looks like) and the objective function on what we have now

• Design the taxonomy collectors actually use, plus golden sets, acceptance criteria, and audit so quality is measurable

• Create the value function for incremental data: given what we have, what is the next example worth?

• Turn that into collection strategy (where, which work, how much, when to stop) and into standards ops and vendors execute against

• Resist false local optima: extra sensors, extra fidelity, extra process that cuts throughput and net information

• Work with marketplace incentives, software, and research so the objective is in the loop — scorecards on coverage, quality, and information, not only volume

• Partner with Evals Lead so dataset decisions and model-capability numbers inform each other

You may be a good fit if you have (Must-have qualifications)

• Quant background (quant research, statistics, decision science, or similar). You think in objective functions and information, not only in label-quality queues

• You can argue about scaling vs fidelity with numbers

• Experience with dataset design, collection strategy, or large-scale data programs

• Fine with in-the-wild collection and a company still scaling countries

Strong candidates may also have experience with (Nice-to-have qualifications)

• Experience at a lab, quant fund, or large-scale data program

• You have designed a taxonomy, golden set, or coverage model used in production

• Familiarity with in-the-wild collection, video, or pose data

• Experience setting quality standards and audit processes that collectors or vendors actually hit

Benefits

• Competitive pay

• Medical, dental, and vision packages with generous premium coverage

• $500 per month credit for waiving medical benefits

• Housing subsidy of $2k per month for those living within walking distance of the office

• Relocation support for those moving to San Francisco (Financial District) or Shenzhen (Nanshan)

• Various wellness benefits covering fitness, mental health, and more

• Daily lunch and dinner in our office

• Unlimited compute budget subject to ROI justification

• Unlimited Codex and Claude credits

• Travel

How we're different

Build believes in the Bitter Lesson. By taking a general approach of learning from humans, our addressable market is all physical labor.

We are a fully in-person team in San Francisco (Financial District) and Shenzhen (Nanshan), and greatly value engineering skills. We do not have boundaries between engineering and research, and we expect all of our technical staff to contribute to both and work across disciplines as needed.

Build AI is an equal opportunity employer. We review every application. If you do not meet every bullet, still apply. Questions: [e-mail]