Job Description
Most AI engineers build agents. Few ever see them run in the real world.
This role is different.
You'll be shipping AI agents into production at real organizations - hospitals, manufacturers, fintechs, real estate firms - and measuring success by the business outcomes those agents generate, not by the demo. You'll be doing it across 3-5 client engagements simultaneously by the end of your first year.
What you're joining
A fast-growing AI consultancy that deploys agentic solutions for enterprise clients across industries. Small team. High standards. Founders still in the work. The kind of shop where your instincts shape delivery approach, your reusable components become the firm's IP, and your track record is real and attributable.
This is not an internal tooling role, a research position, or a "prompt engineer" job with a new title.
What makes this role unusual
You won't be locked into one delivery approach. You'll build custom agents in Python when the problem demands it, and reach for Make, n8n, or equivalent platforms when speed and pragmatism are the right call. The judgment of when to use each is considered as important as the technical ability to execute both.
You'll also be client-facing from day one -- presenting to executives, running discovery on messy workflows, and translating ambiguous business problems into working AI solutions.
What you'll actually be doing
- Owning the full technical delivery of 2 client engagements within your first 90 days, scaling to 3--5 concurrent by month 12
- Building production agents using Python and LangChain/LangGraph (or equivalent) -- not prototypes
- Deploying workflow automation via no-code/low-code tools where appropriate
- Integrating agents into real business systems: CRMs, APIs, internal tools
- Packaging reusable templates, blueprints, and integration patterns that compress delivery time on future engagements
What the role requires
- 3 days a week onsite in Atlanta
- Strong Python -- production-grade, not demo quality
- Hands-on experience with at least one agent orchestration framework
- Comfort building and deploying no-code automations (Make, n8n, Zapier)
- Experience connecting agents to real systems via APIs and function calling
- Working knowledge of RAG, vector databases, and prompt engineering
- Familiarity with at least one cloud AI platform (AWS Bedrock, Azure AI Foundry, or GCP Vertex)
- The ability to talk to a non-technical executive about how an agent works and keep them engaged
Bonus: TypeScript/Node.js, MCP/A2A protocol knowledge, public GitHub presence or published writing on agent development.
What's on offer
Competitive base + performance upside and equity.