The AI platform for precision outbound.
Spray-and-pray outbound is dead. The teams that win do it with precision, and today that means stringing together 10+ tools and hiring a full-time GTM engineer. Cardinal replaces that with agents that build the workflows from natural language: finding the right leads, monitoring what prospects post, writing hyper-specific snippets, and capturing inbound from signups, ads, and site visitors.
Devi Jha
Dropped out of Harvard with(out) a CS degree. 2x YC founder (S23, W26). Sold their last company, Leafpress, to Johnson Controls (NYSE: JCI).
Jianna Liu
MIT CS. Engineering at Meta, DoorDash, and Nvidia. 2x YC founder (S23, W26); Leafpress acquired by Johnson Controls (NYSE: JCI).
Watch the Cardinal launch.
The Cardinal launch · @jianna_liu on X
Agents that do the GTM engineer’s job.
One platform in place of the ten-tool stack: data enrichment, sequencing, and agents that learn from how your team actually sells, all driven by natural language instead of spreadsheets.
The right leads, precisely
Map an entire TAM, enrich a conference list, or build an invite list for a dinner. Cardinal finds the leads worth reaching out to.
Signals, continuously
Agents track what prospects post, share, and engage with, so reps stay current on the people they’re selling to.
Snippets that write themselves
Hyper-specific openers built from what you and your team genuinely have in common with a customer.
Inbound too, not just outbound
Website visitors, product signups, ad impressions, social likes and comments, qualified and routed into sequences automatically.
Very early, already working.
$1M ARR out of the gate
Roughly $1M ARR within weeks of the first launch, with a waitlist behind it, on inbound alone.
40+ YC companies
Running outbound for teams like Mintlify, Greptile, and Luminai, plus a growing set of SF startups.
Second time, together
Leafpress, their energy software company, closed names like Lindt and Honeywell from cold outbound before Johnson Controls acquired it. They are selling the thing they had to build for themselves.
Genuinely early
A months-old company with an exit already behind the founding team. Founding engineers here shape the product, not just the tickets.
Backed by Audacious · Y Combinator (W26). Previously YC S23.
Where the founders show up.
Back at YC, this time on the big stage
“Fast forward to 2026, Jared Friedman and I spoke to more Harvard students about what it’s like starting a company, only this time in San Francisco at the Chase Center. Y Combinator SUS was incredible, the energy was insane.” Read on LinkedIn →
What they’re hiring for.
Founding engineering role in San Francisco, in office five days. This is an application-layer company, so every engineer talks to customers.
Stack: TypeScript, React, Next.js, and Python, on full-stack web architecture with multi-tenant infrastructure, role-based access control, audit logging, and rate limiting. AI/ML experience with LLMs, prompt engineering, and agents is a bonus, not a bar. Olympiad or math-competition background is a plus.
Founding Product Engineer
- Own features end to end across TypeScript, React, Next.js, and Python
- Build LLM workflows and agents that take action across tools
- Data-heavy B2B systems: enrichment, signal detection, real-time monitoring
- Talk to customers directly and make the product calls; design taste matters
- Raw problem-solving and high agency over any particular résumé line
- Some one-off customer work, like custom CRM integrations, but it isn’t the main event
Build the GTM engineer, don’t become one.
In office · San Francisco. Months old, $1M ARR, an exit already behind the founders. Founding engineers get creative freedom and real product ownership.