The first employees at a startup arrive before there is much company to join. The product may be half-built. The job boundaries are soft. A founder is selling a future with a slide deck, a prototype and the nervous energy of someone who has made the leap already. Every accepted offer is therefore two things at once: a hire and a vote of confidence.
Cannon Project has built its business around collecting those votes. The New York firm recruits engineering and go-to-market teams for young software companies, then selectively invests in some of the same businesses. On paper, that makes it a hybrid of search firm, company builder and early-stage venture fund. In practice, its product is judgment: which people are likely to climb quickly, which founders can persuade them to come aboard and which small teams might become durable companies.
Founder Ian Feeney started Cannon Project in 2015 as a technology career agency for high performers, with an early emphasis on athletic and entrepreneurial talent from universities. The public pitch has since widened. Today's website opens with a blunt sentence - "We build founding teams from scratch" - and displays moves into companies such as Ramp, Cursor, Rogo, Decagon, Outtake, Etched and Profound. The before-and-after format is almost comically economical: previous employer, new role, destination. A career condensed into three cells.
The service is also the sensor
Traditional recruiters are paid to fill roles. Traditional venture firms are paid to choose companies. Cannon Project does both, and the overlap is where the model becomes interesting. Recruiting puts the team into live conversations with founders and candidates. It can observe whether a role keeps changing, whether strong operators take the first meeting, whether the founder closes them and whether the people who join stay enthusiastic after the honeymoon.
None of that guarantees an investment return. It does produce a kind of information that is difficult to extract from quarterly metrics. A young company may have little revenue and no mature product, but its ability to attract a respected engineer or an ambitious seller is already visible. Talent movement becomes a rough leading indicator. The firm gets paid for doing the work, develops trust with the people involved and can invest when its conviction rises.
That flywheel also explains the customers. Founders use Cannon Project when a generalist hiring funnel is too noisy and an internal talent department is premature. A three-person company may need its first enterprise seller, an engineer who can tolerate ambiguity or a technical leader who still wants to write code. Candidates use the network from the other side, gaining a guided view of young companies whose quality can otherwise be hard to judge. Limited partners buy the third layer: access to a portfolio selected partly through those operating relationships.
What founders are buying
The core service is early-team recruitment across engineering, product and go-to-market functions. The firm's public placement board ranges from enterprise sales and channel partnerships to operations, threat intelligence and AI deployments. It also lists an executive-search partner and technical advisers, suggesting a model that can reach beyond entry-level matching into leadership and specialized searches.
The problem sounds simple - find good people - but early hiring is a sequence problem. A founder who hires a sales leader before the product is repeatable can create process without demand. A team that waits too long for go-to-market talent may build without learning how customers describe the pain. Engineering candidates, meanwhile, must judge architecture, runway and founder temperament with sparse evidence. Cannon Project's useful role is not merely supplying names. It is matching a company's stage and actual work to a person willing to take that particular risk.
The business model has two engines. Search and placement work can generate fees in the present. Venture funds seek equity appreciation over a much longer horizon. The company does not publish its recruiting fee structure, revenue or valuation, and there is no reason to guess. What is public is the scale of its second investment vehicle: Feeney announced in 2026 that TCP Capital Fund II had closed at $63 million, backed by limited partners including TrueBridge Capital Partners, Hamilton Lane, LGT Capital Partners, Fifth Down Capital and Thrive Capital.
The announcement named more than a dozen companies where Cannon Project had built teams, including Cursor, Ramp, Decagon, Rogo, Avoca, Listen Labs, Profound, Rillet, Baseten, depthfirst, Unify and Outtake. Recent financing announcements supply another view of the investment side. Gumloop named Cannon Project among the participants in its $50 million Series B in March 2026. Unify included the firm in its $40 million Series B alongside Battery Ventures, OpenAI Startup Fund, Thrive, Emergence, Abstract and Capital49.
A category with deliberately blurry edges
Cannon Project occupies a busy intersection. Search firms such as Daversa Partners, Riviera Partners and True Search compete for executive and technical mandates. Embedded recruiters and internal talent teams compete on execution. Accelerators help with formation. Seed funds provide capital and often offer portfolio recruiting. Talent-led investors attempt to make the network itself a sourcing advantage.
The difference is one of order. A conventional fund invests and then helps recruit. Cannon Project often recruits, learns and then decides whether to invest. The service relationship is not an accessory to the fund's pitch; it is meant to generate the access and pattern recognition behind the check. That can be valuable for founders who need people and capital, but it also creates questions a careful customer or candidate should ask. Who is the client in a given search? What information crosses from recruiting into investing? How are conflicts disclosed when the firm has an ownership interest? The hybrid is useful precisely because the functions touch, so the boundaries deserve to be explicit.
There is another limitation in the phrase "high-slope talent." Potential is real, but it is easier to admire than to measure. Prestigious previous employers, university networks and athletic credentials are convenient signals; they can also narrow a pool. Cannon Project's newer public examples show a broader operating range, including experienced salespeople, investors moving into company roles and technical specialists. The test of the model is whether it can identify performance that is not already made obvious by a famous logo on a résumé.
Culture built around the next move
The public team is small enough to fit around one long table. Feeney is managing partner; Andy Linder and Paul De Sadeleer are partners; Sam Warburg leads executive search; Ryan Gross is a principal; and the roster includes advisers across engineering, product and go-to-market work. The firm's own description of the group is "tight-knit." Its visual culture is more clubhouse than corporate campus: community photos, individual career moves and portfolio logos arranged as evidence of a network in motion.
That network is the part competitors cannot quickly copy. A recruiter can buy software and a venture firm can hire talent partners, but neither instantly acquires years of accepted offers, declined offers, alumni moves and founder references. Cannon Project's defensibility depends on whether those relationships remain reciprocal. Candidates must feel that advice serves their careers, not merely a portfolio's headcount plan. Founders must believe the firm understands the job rather than simply circulating the same sought-after names. Limited partners must believe the human signals lead to better selection, not just a busier inbox.
At its best, the model turns a necessary service into a learning system. Every search updates a map of which skills are scarce, which companies are gaining pull and which operators are ready for more responsibility. Every successful placement makes the network more useful. Every investment creates a longer relationship with a team Cannon Project may already know seat by seat.
Where the wager lands
Cannon Project fits most naturally in pre-seed and seed-stage B2B software, especially where technical density and early distribution matter. Its named companies cluster around AI infrastructure, developer tools, enterprise software and modern go-to-market systems. These markets move quickly, compete for overlapping talent and often need small teams to carry disproportionate weight. A recruiting-led investor can plausibly see shifts before they appear in a financing announcement.
The firm's $63 million second fund gives the strategy more room, but it does not change the basic bet. Cannon Project is wagering that the work of building teams will continue to produce better relationships and earlier insight than watching from the cap-table sidelines. Founders can use it as a concentrated hiring partner with the possibility of capital. Candidates can use it as a guide to consequential roles that are difficult to evaluate from job listings alone. Investors can use it as a window into the labor market beneath the startup market.
The amusing truth is that venture capital likes to describe itself in grand abstractions, while companies are built through specific calendar problems: the candidate who needs one more conversation, the reference who calls back late, the offer that expires Friday. Cannon Project has chosen those mundane moments as its point of entry. The check comes later. First, somebody has to take the job.
Explore Cannon Project
Follow the firm's current work, team and portfolio through its own channels, or read recent announcements from companies it has backed.