Customer Engineer

About Cardinal

Cardinal (trycardinal.ai) is the AI-native outbound sales platform replacing the SDR-and-GTM-engineer stack that companies cobble together today with Apollo, Outreach, Clay, and a few engineers gluing them all together.

We're a YC W26 company. In our first few months we've grown to $80K+ MRR, with 50+ YC companies and a growing roster of post-Series-A startups running their outbound on us. We've raised over $5M from Audacious, Burst Capital, and YC, and we're targeting $10M ARR by mid-2026.

We're a small, dense, design-obsessed team. The bar for engineering taste, product instinct, and aesthetic quality is very high — if you've ever opened Linear or Attio and thought "why doesn't every B2B tool feel like this," that's the bar.

The Role

Our customers are founders and GTM leaders who have handed us their pipeline. When something breaks, or a sequence doesn't behave the way they expected, or they need a field we don't support yet, they say so in a shared Slack channel — and how fast we turn that around is most of what determines whether they stay.

As a Customer Engineer, that loop is yours. You're the first responder in every customer channel and the person who actually fixes the thing. Not triage-and-hand-off: you reproduce it, find it, and ship the fix, or you write the one-off script that unblocks them today while the real fix gets specced.

You'll sit on onboarding and check-in calls, so you hear the friction firsthand instead of through a ticket. You'll get customers to their first launched campaign and keep them sending. And you'll be the highest-signal input our product roadmap has, because you're the only person who sees every request, every workaround, and every place the product quietly confuses people.

Some companies call this a support engineer or a forward-deployed engineer. The distinction that matters to us: you have write access and you use it.

What you'll do

• Own first response in customer Slack channels — reproduce, diagnose, and give a real answer, fast
• Ship fixes directly: bugs on live campaigns, data and enrichment issues, sequence and sender misconfigurations, backfills, one-off scripts
• Decide what's a same-day fix versus what needs a spec, and write the spec when it's the latter
• Join onboarding and weekly check-in calls; unblock customers live rather than taking notes
• Build the internal tooling that makes the next fix faster — admin views, replay and debugging tools, health checks that catch problems before the customer does
• Turn recurring friction into evidence-backed tickets so the product stops generating the same requests

We'd love to meet you if you

Philosophy: You don't file the bug, you fix it. A customer waiting on you is a thing you feel. You move fast without confusing fast for sloppy, and you know which of the two a given situation calls for.

Debugging range: You're comfortable dropping into unfamiliar code and following a problem wherever it goes — frontend state, API, queue, third-party integration, someone else's rate limit. You're fluent in logs, traces, and production data.

Customer instinct: You can tell the difference between what a customer asked for and what they actually need, and you can say so without being difficult about it. You write clearly under pressure, including when the answer is "this is broken, here's when it'll be fixed."

GTM context: You've worked at a sales-adjacent company — an outbound tool, a CRM, a sales intelligence platform, a revenue-tech startup — or you've been the technical person at a startup who actually built and ran the outbound motion yourself. You should immediately understand why a "step 3 reply detection" bug or a "lead enrichment timeout" matters to a customer. If words like sequence, enrollment, ICP, intent signal, deliverability, and sender rotation don't make you glaze over, we want to talk.

Bonus points if you

• Have done forward-deployed, solutions, or support engineering at a technical product company — and have opinions about how to do it without becoming a ticket queue
• Have founded a company, been an early engineer at a YC-stage startup, or shipped a product 0→1
• Have built or worked at a GTM/sales/CRM/enrichment tool — Apollo, Outreach, Clay, Attio, HubSpot, Salesforce, Gong, Common Room, or anything in that orbit
• Are a power user of GTM tools yourself — you've actually run outbound, written sequences, and fought with deliverability
• Have built internal tooling or observability that other people ended up depending on

This is an in-person role at our SF office. We're early stage, which means working hard and moving quickly. Apply if that excites you.

About Cardinal

Cardinal is the AI platform for precision outbound. We run outbound for 40+ YC companies, including Mintlify, Greptile, and Luminai.

"Spray and pray" outbound is dead. The best sales teams operate with extreme precision — finding things in common with customers, monitoring their content, mapping inbound signals, and tracking website visitors. But doing this well means stringing together 10+ tools and hiring a GTM engineer. Cardinal replaces all of that.

Our AI agents find hyper-targeted leads, continuously monitor what prospects are posting and engaging with on social media, generate personalized snippets based on shared context, and orchestrate end-to-end outbound workflows across email and LinkedIn. Cardinal also captures inbound from website visitors, product signups, ad impressions, and social engagement — qualifying and routing them into sequences automatically.

Cardinal is founded by Devi and Jianna, 2x YC founders (S23, W26) and Harvard/MIT CS alums. Their last company, Leafpress, was an energy software company where they closed customers like Lindt and Honeywell from cold outbound.