Senior AI Agentic Engineer to $5000

Senior Agentic AI Engineer

Rozeta Labs · Full-time Contractor · 100% Remote · $4,000 – $5,000 / month · EST overlap preferred

About Rozeta Labs

Rozeta Labs is an early-stage AI implementation company. We design and build custom AI automation systems for businesses that have real operational problems — home services, professional services, health and wellness, private equity-backed companies, and beyond. Our clients are owner-operators and leadership teams who need AI to actually work inside their business, not just demo well.

We don't sell software. We sell outcomes. We go inside a client's operation, identify where AI creates the most leverage, build it, and make sure it runs.

We're a startup. Small team, high standards, moving fast. This is a foundational hire and you'll work directly with the founder on every engagement.

The Role

You're the technical engine behind every client engagement. When we bring on a new client, you design the AI system, build it, maintain it, and expand it over time. You'll sit directly alongside the founder — shaping how we architect solutions, how we deliver them, and how Rozeta evolves as the AI landscape shifts under our feet.

This is not a research role. Not a prompt engineering role. Not an internal tooling role. You build production AI systems that drive measurable business outcomes for real companies — and you've done it before.

As we grow, you'll be responsible for hiring additional engineers who will help with an increase in client engagements.

What You'll Build

• Multi-agent orchestration systems that automate complex business workflows end-to-end
• RAG pipelines that give AI agents access to client-specific knowledge — pricing, inventory, history, policies
• Tool-use architectures that connect AI to CRMs, scheduling platforms, and business software via APIs and MCP
• Human-in-the-loop workflows with deliberate escalation logic and clean handoffs
• Evaluation frameworks that actually measure whether a system is working

What We're Looking For

• Direct, hands-on experience implementing AI systems that drove real business outcomes for small, mid-size, or enterprise clients — not internal tools, not side projects, not demos
• Deep fluency with agentic AI architecture: multi-agent patterns, planner/executor design, tool use, failure handling, observability
• Extensive experience with the Anthropic SDK and Claude API — tool use, extended thinking, long context, MCP
• Hands-on experience building with AI-native coding tools: Claude Code, Codex, Cursor, and similar agentic development environments
• Strong Python; hands-on experience with LangGraph, LangChain, or equivalent orchestration frameworks
• Real RAG experience: embedding models, vector databases, retrieval quality — not just the concept
• The ability to look at a business process and translate it into an agent workflow — you think in systems, not just in code
• Strong written and spoken English — you'll communicate directly with US-based founders and clients

• Comfort working in a fast-moving startup where the scope expands and the playbook is still being written

   

Nice to Have

• Background in solutions engineering, technical consulting, or client-facing implementation roles
• Experience with automation platforms like Zapier or Make in a technical capacity
• Familiarity with business software across verticals — CRMs, scheduling tools, EHRs, field service platforms

• Experience building and shipping AI products or services, not just internal automation

   

Compensation and Structure

• $4,000 – $5,000 / month depending on experience
• Full-time independent contractor, paid monthly
• 100% remote, async-first with regular overlap during EST business hours

• You'll go through a technical evaluation as part of the interview process

   

To Apply

1. Submit Resume

2. Record a short Loom (3–5 minutes, no editing required).

Tell us who you are and walk us through the most interesting AI system you've ever built. What did it do, who did it help, and how do you know it actually worked? We want to see how you think — not a polished presentation.

Applications without a Loom will not be reviewed.