AI Agent Builder for Sales

We use AI every day to help clients win — now we want to do the same for ourselves. We're hiring someone who can stand up real, in-production sales agents: not demos, not weekend hacks. Agents that research accounts, qualify leads, draft outreach, run follow-ups, and keep the CRM honest. You'll work side by side with the founder and our sales lead. The bar is high — agents we actually ship, with guardrails, evals, and measurable lift on pipeline.

• Design and ship AI agents for sales workflows — account research, lead qualification, personalized outreach drafting, follow-up sequencing, meeting prep

• Build retrieval pipelines over our past proposals, case studies, and CRM history so agents speak in DoodleWeb's voice and reference real wins

• Integrate with the tools we already use — HubSpot/Pipedrive-class CRMs, LinkedIn, email, Slack, calendar, and our website's lead intake

• Set up evals, guardrails, and human-in-the-loop checkpoints so agents fail safely

• Instrument everything — track agent actions, win-rate lift, response rate lift, time saved per rep

• Iterate fast with the sales team: agents that don't improve weekly get killed

• 3+ years building production software; 1+ year shipping LLM-powered agents or workflows that real users depend on

• Strong TypeScript or Python; comfortable with frameworks like LangGraph, Mastra, Vercel AI SDK, or your own orchestration

• Hands-on with OpenAI, Anthropic, Gemini, and open models — and opinions about when to use which

• Experience with RAG, vector stores (pgvector, Pinecone, Turbopuffer), and prompt evaluation

• Comfort wiring up integrations (REST, webhooks, OAuth) and writing your own internal tools when something doesn't exist

• Product instincts — you measure agents by business outcome, not token count

• Experience selling to or working inside an agency or B2B services company

• Built voice agents (Vapi, Retell, LiveKit) for outbound or qualification calls

• Familiarity with MCP servers and tool-use protocols

• Background in deliverability — domains, warmup, SPF/DKIM/DMARC

• Shipped agentic features in a public product

• Competitive base salary or contract rate

• Performance bonus tied to qualified pipeline created by agents you built

• Full remote work

• Generous AI tooling and compute budget (every model, every API)

• PTO, health benefits (for full-time), and conference budget

1. 1

   Application review (within 72 hours)

2. 2

   30-minute intro call

3. 3

   Paid build: one real agent, flat fee, 1-week turnaround

4. 4

   Architecture and code review

5. 5

   Final call and offer