Job Description & Details
This is a hands-on 6-month contract role where you will be building and scaling production-ready AI agents and LLM chatbots on Google Cloud Platform. If you are tired of just pushing out superficial prompt wrappers and want to tackle actual multi-step reasoning, tool orchestration, and infrastructure deployment, this gig in Atlanta is worth a serious look.
What You'll Actually Be Doing
You will own the lifecycle of AI solutions from rapid proof-of-concept to production deployment on GCP services like Cloud Run and GKE. Day-to-day, expect to write robust Python code for agent frameworks, wire up complex APIs and databases, handle vector stores, and ensure your autonomous workflows don't hallucinate or leak data. You'll also be dealing with the unglamorous side of AI engineering—optimizing inference latency, managing cloud costs, and setting up solid monitoring and guardrails.
The Core Tech Stack
Python is non-negotiable here, paired with hands-on experience using Google Cloud Platform (especially Vertex AI and BigQuery) and agent development kits or modern agent frameworks. You'll need to know your way around distributed systems and APIs because these agents aren't living in isolation—they need to securely talk to enterprise databases, internal microservices, and external tools without breaking production.
Interview Expectations
Expect the hiring team to press you on how you handle state management and context windows in multi-step agentic workflows, likely asking you to design a fallback strategy when an LLM fails to call a tool correctly. They are secretly testing whether you understand production failure modes or if you've only built local toy demos. You should also be ready to explain how you would architect a secure RAG pipeline that enforces strict IAM policies and data access controls against a massive BigQuery dataset.
Application Advice
Make sure your resume screams GCP and agent orchestration rather than just generic full-stack development. Explicitly highlight any experience you have taking LLM applications from a local Jupyter notebook or Agent Studio prototype into a deployed, monitored container on Cloud Run or GKE. If you've dealt with supply chain data, vector databases, or complex multi-agent architectures, put those front and center to clear the ATS and catch the engineering manager's eye.