Instantly
Apply now

Instantly AI team · remote · hiring

Build agentsthat sellovernight.

We are the small team shipping the AI behind outbound for 45,000+ businesses. Production LLMs, real users, your work live in days.

Apply nowScroll the pipeline

One lead, end to end

This is the loop our agents run millions of times. You would own a piece of it. Scroll to follow one lead through.

  1. 01 / 05

    Find the right buyer

    Agentic search across a 450M+ contact database turns a fuzzy ICP into a ranked list of real people.

    agent.run(lead_8214)
    • search.icp("RevOps leads, B2B SaaS, 50-250")
    • scanned 450M+ profiles
    • ranked 1,284 matches

  2. 02 / 05

    Research every account

    The agent reads the company, the role and recent signals, then decides what is worth saying.

    agent.run(lead_8214)
    • fetch.company_signals()
    • hiring: 3 SDR roles open
    • angle: ramp new reps faster

  3. 03 / 05

    Write like a human

    Personalized outreach drafted per lead, with guardrails, evals and the sender’s own voice.

    agent.run(lead_8214)
    • draft.email(lead, angle)
    • eval: tone ok · facts grounded
    • subject: “3 new SDRs?”

  4. 04 / 05

    Land in the inbox

    Sending across warmed accounts with deliverability checks, so the work actually gets read.

    agent.run(lead_8214)
    • schedule.send(window="09:14 local")
    • inbox placement check passed
    • delivered

  5. 05 / 05

    Handle the reply

    Replies are read, classified and answered with a draft, then routed to the right next step.

    agent.run(lead_8214)
    • classify.reply()
    • intent: interested
    • draft reply ready for review

We build heavily with AI. Less time typing, more time thinking.

- how the AI team ships

The stack, no buzzwords

Production LLMs on a modern TypeScript stack, at the scale of hundreds of millions of lead profiles.

AI and agents
LLMs (OpenAI and more),Vercel AI SDK,Agents and tool calling,MCP,Evals and tracing,RAG
Product
TypeScript,React,Node.js
Data
PostgreSQL,Elasticsearch,Redis,MongoDB
Infrastructure
AWS Lambda and SQS,Cloudflare,Datadog,LaunchDarkly
How we build
AI coding assistants,Feature flags,Ship daily

Tools evolve with the work. Bring the ones that make you faster.

How the AI team works

Ownership end to end

You take features from idea to production and they stay yours. No hand-offs, no ticket factories.

Remote across 15+ countries

Work from anywhere alongside 50+ remote colleagues. We write things down and respect time zones.

Speed with judgment

We move fast without breaking things. Systems that work beat ideas on a slide.

Builders hire builders

Every application is reviewed by a founder, and engineers you would work with run the interviews.

  • Clear growth paths and mentorship
  • Generous time off and mental health days
  • Premium software and AI tools

How hiring works

  1. 01

    Send your work

    Email your CV plus links to AI systems you have shipped: a repo, a demo or a write-up.

  2. 02

    Founder review

    A founder reads every application. If there is a fit, we reach out.

  3. 03

    Technical conversation

    Talk systems, agents and trade-offs with the engineers you would work with.

  4. 04

    Build something real

    A practical exercise close to the work, so you see the job and we see how you build.

You will fit in if you

  • Have built and shipped AI systems recently
  • Understand LLMs, agents or workflows deeply
  • Care about execution more than theory
  • Like owning a system end to end

Probably not for you if you

  • Are looking to learn the basics
  • Have never deployed a real system
  • Want a big-company process
Made with Instantly