Job Description & Details
This is a heavy-duty enterprise AI engineering role where you aren't just writing toy scripts, but actually building production-ready agentic workflows and ML pipelines on AWS. If you enjoy wrangling complex LLM architectures and moving models from messy notebooks into scalable cloud infrastructure, this gig is worth your attention.
What You'll Actually Be Doing
Expect to spend your days knee-deep in Python 3.11, architecting RAG pipelines, and wiring up multi-agent frameworks like LangGraph or CrewAI to talk to various enterprise APIs and vector databases. You will own the full model lifecycle—from data prep in tools like Dataiku to deployment, monitoring, and cost optimization on Amazon SageMaker and AWS Bedrock. The real challenge here is bridging the gap between experimental AI prototypes and secure, reliable production systems that enterprise stakeholders actually trust.
The Core Tech Stack
Python is non-negotiable here, and you need to know it inside out alongside modern async patterns and API integrations. On the cloud side, deep fluency in Amazon SageMaker and AWS Bedrock is mandatory, coupled with hands-on experience orchestrating LLMs using frameworks like LangGraph, LlamaIndex, or AutoGen. Knowing your way around vector databases, MLOps tooling, and Dataiku will separate you from the rest of the applicant pool.
Interview Expectations
Be ready for a deep dive into how you handle state management and failure recovery when building multi-agent AI workflows. The hiring manager is going to test your system design chops by asking how you scale RAG pipelines and optimize inference costs on AWS without sacrificing latency. They want to see if you can debug complex asynchronous Python code on the fly and explain how you handle model drift in production.
Application Advice
Make sure your resume doesn't just list tools, but highlights specific instances where you took an LLM or ML model all the way to production. Explicitly call out keywords like Amazon SageMaker, AWS Bedrock, LangGraph, RAG, and Python 3.11 so you easily clear the ATS filters. Emphasize your MLOps background and any enterprise-scale deployment experience you have to prove you aren't just a prompt engineer.