Across the application layer, every investment we make is, in some way, a bet on AI delivering ROI. The organizing principle for those investments is that the marriage of proprietary data and frontier intelligence can produce compounding ROI for customers.
We look for founders who understand a powerful learning loop. A company does important work on behalf of its customers and, in the process, collects proprietary data and context. It exposes that data to increasingly capable intelligence, allowing it to automate better business decisions. Those decisions generate more data, context, and feedback, making the system more capable and the ROI more substantial over time.
This learning loop is at play throughout our portfolio. Tabs is reinventing revenue management, Inspiren is transforming elder care, and Casap is changing how companies manage disputes and chargebacks. They serve very different markets, but the underlying architecture is similar. Each company is building a proprietary data advantage, connecting it to frontier intelligence, and using the combination to automate increasingly consequential decisions for its customers.
Over the past year, we have been applying the same principle internally. We assembled a team of four and began building PrimaryOS. Our honest assessment is that the progress has been real, but modest. We have automated workflows such as outbound founder campaigns, improved parts of our diligence, and developed better portfolio views. These are useful gains, but they do not yet amount to transformation.
At the same time, the work has made us more convinced than ever that transformation is possible. We believe we can generate meaningful alpha by betting harder on compounding systems of intelligence. We are beginning to connect Primary’s proprietary data, institutional knowledge, and daily work to frontier models, while giving people across the firm the tools to build and experiment for themselves.
This is consistent with how we have always built Primary. We think of ourselves as builders, not backers. We built the largest portfolio support team of any early stage venture firm. Primary Labs is our incubation engine, producing six or more new businesses a year. Primary Summit has become one of the largest annual gatherings of the New York technology ecosystem. We believe the best venture firms create capabilities, not just deploy capital.
PrimaryOS is the most ambitious capability we have undertaken inside the firm. We have learned enough over the past year to see its potential, and now we want to double down. We are hiring a Head of AI to help us completely reinvent how Primary works for the age of AI.
The Ambition
Venture capital can be reduced to four jobs: find, decide, win, and exit. The words are simple. The work underneath them is confoundingly complex.
Finding an exceptional founder early requires enormous surface area and strong pattern recognition. Developing conviction requires making decisions with incomplete evidence and understanding a market before its future is obvious. Winning requires trust and an understanding of what a particular founder needs. Helping a company succeed requires knowing which person, customer, or intervention could change its trajectory.
Our objective is alpha: finding exceptional companies earlier, building sharper conviction, winning the investments that matter, and helping those companies become more consequential.
Achieving that will require us to reconsider the workflows themselves. What should sourcing look like when an intelligent system can evaluate emerging founders and learn from our decisions? How should we research a market when agents can assemble evidence, challenge a thesis, and update our view? How does portfolio support change when we can match the needs of hundreds of companies against thousands of people, customers, and resources with far greater precision?
These questions extend across investing, portfolio support, finance, LP relations, incubation, and firm operations. The opportunity encompasses everything we do.
The Role
The Head of AI will own the strategy, architecture, product, and team behind PrimaryOS. You will work directly with our Chief Investment Officers and leaders across Primary to define what AI should make possible for the firm, decide what to build and in what order, and measure whether it materially improves our work.
The technical mandate includes our context and data layer, models, retrieval, agents, evals, permissions, shared tooling, and feedback systems. You will also own the application layer, building products that improve how we find companies, conduct diligence, make decisions, win investments, support founders, and run the firm.
You will lead the PrimaryOS team and grow it over time. This is a leadership role, but we are looking for someone who still wants to build. You should be comfortable writing code, prototyping, reviewing architecture, assessing vendors, and getting directly involved when the problem demands it.
The organizational challenge is as important as the technical one. Primary is a complex firm with sixty smart, opinionated people working across investing, recruiting, go to market, finance, networks, incubation, and firm operations. You will need to understand their work, identify common problems beneath different workflows, and earn enough trust to change how people operate. You will also need to bring coherence to the experimentation already happening without diminishing its energy.
Who You Are
There is no single obvious background for this role. You are likely some combination of engineering leader, product thinker, and founder, with meaningful experience building complex systems that people actually use.
You have worked on AI products that advance knowledge work and have seen what it takes to move from an impressive demonstration to a system that changes how people research, reason, decide, and collaborate. You are current on the technical frontier, with hands on fluency across LLMs, agents, retrieval, evals, data infrastructure, and the rapidly changing model ecosystem.
You have also operated inside a complex organization. You understand that stakeholder management is part of the product, particularly when users have different workflows, incentives, and levels of technical fluency. You can listen carefully, form an independent view, and build alignment among smart people without sanding down the ambition.
Most importantly, you want something entrepreneurial. This role comes with an existing team, proprietary data, committed users, direct access to the firm’s leadership, and a substantial surface area of meaningful problems. It does not come with a predetermined roadmap. You will have enormous freedom to determine what we build, how we build it, and what an AI native venture firm should become.