Head of R&D (Founding Team)

About the role

We're hiring the person who will own R&D at Yolk - engineering, QA, and applied AI research - as a founding-team hire.

That means all of it: what we build and how, the quality bar and how we enforce it, the AI architecture and the research that keeps it ahead, who we hire, and when we're ready to ship. You'll inherit all three functions, decide what they should look like at 3x, and build that.

You'll report directly to the founder and work alongside our Lead PM, who owns product direction and roadmap. You own how it gets built, how good it is, and how fast it moves.

A word of honesty

This is a founding-team role at an early-stage company, which means it is hands-on. You will get into the code. You will run the QA passes yourself, on the things that matter most, even though QA isn't your only job. You will test a video session on a bad connection at 9pm because a launch depends on it. You will make calls with incomplete information and live with them. You'll lead R&D, but you'll lead by doing, not by directing. If you want a purely strategic seat with people beneath you to execute, this is the wrong role. The reverse is just as true: this is not a senior engineering job with a bigger title on it. You'll be managing people, hiring them, and answering for their work, and there's no technical leader above you to learn that from. If you haven't done it before, this isn't the place to start. If you want to own the hardest, most consequential technical problems in the company and see them through, read on.

What you'll own

• Reliability of realtime AI agents, audio and chat sessions on a user's old laptop and a sales rep's poor wifi, not just a developer's machine.
• Turning a launch-phase codebase into one a growing team can work in safely.
• The quality bar for everything that ships. We have a QA function today, not a QA department — part of this role is deciding what it should become, and how much of it is people versus automated test infrastructure. We don't have a fixed answer; you'll set it.
• End-to-end quality of the booking-and-payments flow, including the failure paths most teams skip.
• Security and privacy of sensitive sales conversation data — the controls that keep us defensible, and an independent security review you can run yourself.
• The AI and agent architecture: what models, what orchestration, what runs on-device versus in the cloud, and how we stay ahead as the underlying models move.
• Evaluation as a discipline — you own how we know whether the coaching is actually getting better. The evals, the benchmarks and the harnesses that tell us a change is real and not a vibe, on a product where the output is a judgment call rather than a right answer.
• Turning research results into shipped product, not internal demos.
• Hiring, structure and standards across all three functions.
• A delivery cadence the company can plan around.

What we're looking for

• You've managed engineers directly for 2+ years — as their manager, not as a tech lead with influence. You've hired, set the bar, run the uncomfortable conversations, and owned the outcome when someone wasn't working out.
• 8+ years building software overall, and you've owned technical direction rather than executing someone else's architecture.
• Hands-on operator first, leader second: you lead through doing. That does not mean this is a senior IC role with a title on top — you'll be the only technical leader in the company.
• You've shipped consumer-grade B2B software to production and through real users, on something that affected your customers' revenue.
• Real depth in at least one of: realtime/low-latency systems, or applied LLM and agent systems in production. Credible curiosity in the other.
• You can run a security review and stand up the controls behind it yourself, not just brief someone else to.
• You work directly and candidly with a non-technical founder, and you translate technical decisions for non-technical stakeholders without dumbing them down.
• You've worked somewhere where the product and user experience are real. You're not learning why it matters here.
• Strong communicator who can translate technical decisions for non-technical stakeholders.
• You have a real opinion about quality — what to automate, what to test by hand, and what to knowingly let slide before a launch.

Nice to have

• You've been part of an early-stage or founding team before and know how to operate without much process.
• You've owned a QA or quality function, not just written tests.
• You've run evals or benchmarking for an LLM product in production.
• You use AI coding tools seriously in your own work and have views on what that does to a team's throughput and its review standards.

What success looks like — first weeks

• Own and close out the post-launch readiness so we can scale; get Yolk going in its first market with confidence, not crossed fingers.
• Prove the full AI signal-to-won-deal flow end-to-end, including every failure path, across real sales calls and real deals.
• Give us an honest read on where the platform actually stands: what's solid, what's held together with tape, and what you'd fix first.

What success looks like — first 6–12 months

• Turn a launch-phase-built product into a maintainable, well-owned codebase with a team to match.
• A quality system that catches regressions before customers do, and that you can show us the numbers on — plus a clear answer on what the QA function should grow into.
• An AI architecture that improves measurably, with evals to prove it — not just changes that feel better.
• Establish a sane delivery cadence and the engineering and R&D processes we'll scale on, and the hires to run them

Why this role matters

In a category built on trust, the technical decisions are the business. Whether a session holds, whether the platform feels solid — that's what earns the trust everything else depends on. As a founding-team member you'll build the foundation Yolk grows on, and shape the company from its first days live.

What we offer

A founding-team seat reporting directly to the founder, with real authority over all of R&D. A product that's already built and a launch to drive, not a blank page. Competitive cash compensation plus meaningful equity, reflecting the seniority and ownership of the role.

About Yolk

A sales rep tracks nine things at once on a live call. Human working memory holds four. Every AI sales tool tries to add more dashboards on top. We subtract — Yolk's AI coach works inside the live call, surfacing the one right thing to say at the moment it matters, then turns what broke down on the call into targeted AI roleplay practice afterward.

The facts: launched end of May 2026. Hundreds of salespeople coached in the first weeks, across several live pilots and paying clients. Thousands of real sales calls already analyzed. Backed by investors in Anthropic and Groq. SOC 2 compliant, 4 patents in core AI methods.