Artificial Intelligence Engineer

• Build production -grade agentic systems on AWS and OBZ's Neo platform: multi -agent orchestration, RAG pipelines, tool design, MCP server development, human -in -the -loop workflows, and the integration layer that connects agents to real enterprise systems with real data quality constraints and real security perimeters.
  
• Write evaluations before you write features. Define what good performance looks like, build the harness that measures it, and test against it throughout the build. In regulated environments, systems that have not been evaluated do not reach production.
  
• Integrate into enterprise reality. Legacy APIs, inconsistent data availability, access control constraints, and compliance requirements that were not designed for agentic systems. Build within those realities, not around them.
  
• Review and own what you build through production. Clean code, tested code, documented code. If the system degrades in production, you investigate. If the architecture needs to change, you contribute to the decision. You are accountable for what you ship.
  
• Practice DevSecOps from the first line of code. CI/CD, infrastructure as code, secrets management, IAM discipline, and observability are baseline practices. In regulated deployments, security is an architectural input from day one.
  
• Contribute reusable IP at every engagement: code modules, evaluation frameworks, deployment runbooks, agent design patterns. Every engagement leaves something the next squad can build on.
  
• Stay at the frontier. Know what AWS releases across its AI and agentic services stack. Know what the leading AI labs are publishing about production agentic systems: Anthropic, OpenAI, Google DeepMind, Meta, and the open -source frontier. Know what breaks in production and why. Bring that knowledge into your engagements before clients ask.
  
• Engage with client technical teams. Participate in design reviews, explain what you built when asked, and represent your work credibly in the room.
  
  

Requirements

• Strong programming fundamentals in Python, with production -grade code discipline. Additional language experience in TypeScript, Java, or Go is a plus.
  
• Genuine agentic AI fluency. You have built with agent frameworks: LangGraph, LangChain, CrewAI, or equivalent. You understand RAG architectures, vector databases, prompt engineering at scale, evaluation harnesses, and the failure modes specific to agentic systems in production. At junior levels this can be demonstrated through personal projects and self -directed learning. At senior levels it must be demonstrated through production systems.
  
• AWS proficiency. You have built on AWS. You are working toward or hold AWS certifications relevant to your level, including AWS AI Practitioner, AWS Certified Machine Learning Engineer Associate, or AWS Solutions Architect. Anthropic certification is also valued.
  
• DevSecOps fluency. CI/CD, testing, infrastructure as code, and security hygiene are not optional. You know what a pull request review is for and you take it seriously.
  
• A strong academic foundation in computer science, engineering, mathematics, or physics. For recent graduates, this is the primary signal of aptitude we look for alongside evidence of self -directed building.
  
• Curiosity and ownership. You read what the labs publish. You experiment with what they release. You find the problem interesting. When something is broken you own the fix.
  
• Client -facing communication. You can explain what you built to a non -technical stakeholder with precision and without jargon. You participate in technical reviews with composure and represent your work credibly in the room.
  

• You have built a production agentic system that a regulated enterprise depends on, and you can walk through exactly what it does, how you evaluated it, where it failed in testing, and what you did about it.
  
• You have identified a production failure mode before it reached the client environment and fixed it under time pressure without anyone asking you to.
  
• You use AI coworkers to accelerate your engineering work and you have a clear point of view on which tasks they handle well and which ones still require your direct judgment.
  
• You have contributed a reusable technical asset that another squad adopted and you know which engagement is using it.
  

• The client has a production agentic system that is performing, that their team understands and can operate, and that is producing the outcome that justified the investment.
  
• OBZ has something reusable from every engagement you touch: a design pattern, an evaluation framework, a deployment runbook, a code module that the next squad picks up without starting from scratch. Your work compounds across the firm, not just within a single account.
  
• You are a more capable engineer at the end of every engagement than you were at the start. You have pushed into something you had not done before, solved a problem that was not in the playbook, and left evidence of that growth in the IP you contributed and in the engineers you developed alongside you.
  
• When all three of these are consistently true, you are succeeding at this role.
  

• AWS AI Practitioner, AWS Certified Machine Learning Engineer Associate, AWS Certified Solutions Architect (Associate or Professional), and Anthropic certification are all relevant and valued. Certification expectations are calibrated to level and assessed during the hiring process.
  

Benefits