About AI Circle

AI Circle is building the global network for the people shaping the future of AI.

We bring together exceptional founders, researchers, engineers, executives, and AI professionals to foster meaningful connections, exchange ideas, and create opportunities across the rapidly evolving AI ecosystem. Through a curated community, events, and talent initiatives, we're creating a place where the best minds in AI can connect and collaborate.

The Opportunity

Frontier AI models need more than public internet data—they need the tacit expertise and judgment of professionals with real-world experience.

Our client is a newly launched human data business built to bridge this gap for frontier AI researchers. Backed by one of the world’s largest expert networks, it has immediate access to millions of vetted experts across highly specialized domains, supported by established governance, compliance, and confidentiality infrastructure.

The business combines startup speed with enterprise-scale expertise and trust, creating high-impact opportunities for senior operators.

About the role

You will own how expert data actually gets built and shipped. That means running the earliest programs with your own hands, then converting what you learn into the systems, playbooks, and team that let the company run many at once.

This is explicitly a player-coach role, and the balance shifts over time. In the early months you are in the work: scoping directly with researchers, writing the first annotation guidelines and rubrics yourself, personally recruiting and calibrating the first cohort of experts, sitting in the war room when a program is running hot, and diagnosing why throughput or quality is off. There is no existing playbook to inherit and no team to delegate to yet.

As volume grows, you become the coach codifying what worked into repeatable process, hiring program managers and quality leads, and training them to hold the standard you set. The systems you build in month three are what a team of ten runs in month twelve. Candidates who only want to design the operating model, or who only want to run programs, will not enjoy this. The role is both, in that order.

You will be the connective tissue between researchers who need a specific kind of data, experts who produce it, and a business that has to deliver on time and at a margin that works.

What You'll Own

Delivery, end to end

From scope of work through accepted final delivery and renewal. You set milestones, staffing plans, and acceptance criteria across supervised fine-tuning data, preference and reward-model data, red-teaming, and evaluation and benchmark construction. When a program is at risk, you run the daily cadence that gets it back on track.

Quality

Define what “good” means for every program and build the system that enforces it: annotation guidelines, rubrics, calibration sessions, gold datasets, inter-annotator agreement tracking, and tiered review. The central challenge of this job is turning a researcher's underspecified rubric into instructions that a hundred experts interpret the same way.

The expert cohort

Convert the network into production-ready panels: screening, onboarding, training, calibration, performance management, and retention. These are accomplished professionals with limited patience and real day jobs, so engagement is an operating discipline, not an afterthought. In this market, the cohort you can hold onto is the moat.

Throughput, cost, and margin

Own contributor cost, utilization, and rework. Diagnose bottlenecks and restructure workflows, instructions, incentives, and review layers to move quality and throughput at the same time. Build the cost model for new program types, and know when a deal is priced to lose before it is signed.

The client relationship on delivery

Be the person researchers trust to tell them the truth about scope, timing, and what the data will and will not support. Own delivery reporting and the technical credibility that earns the second program.

The team and the operating system

Hire, train, and lead program managers and quality leads. Turn your own reps into documented process, then hold the bar as other people run it. Growing this function is the deliverable, not a side effect.

What We're Looking For

Minimum qualifications

• 8+ years of experience in delivery, program, or operations roles, including ownership of complex client programs against contractual milestones.

• Direct experience in human data or AI training data annotation, evaluation, preference data, red-teaming, or comparable programs built by distributed human contributors. This is a requirement, not a nice-to-have.

• You have owned the quality of work produced by people you did not directly manage, and can explain the system you used to do it.

• Evidence of building process where none existed, and of hiring and training the people who then ran it.

• Commercial fluency, you have owned or materially influenced margin, cost, or utilization on services work.

Preferred qualifications

• Working fluency in supervised fine-tuning, reinforcement learning from human feedback, preference and reward-model data, and model evaluation, and comfort in a room with researchers.

• Experience delivering into frontier labs or enterprise AI research teams.

• Existing relationships with researchers, data operations teams, or procurement at frontier labs or enterprise AI organizations.

• Depth in regulated or high-expertise domains such as healthcare, legal, or financial services.

• Experience standing up a delivery or quality function at an early-stage company.

You'll thrive here if

• You have opinions about why programs fail. You have watched one go sideways because the guidelines were ambiguous, the cohort burned out, or someone accepted a scope nobody could staff.

• You are genuinely willing to do the work before you have a team, and genuinely good at handing it off once you do.

• You would rather define the standard than inherit one, and ambiguity reads as opportunity rather than risk.

Compensation and Benefits

Base salary range: $250,000 - $300,000. The final offer depends on experience and level. The role also includes a performance bonus and equity participation.

Benefits include medical, dental, and vision coverage, a 401(k) with company match, and paid time off.