Artificial Intelligence Engineer

Role Overview

As an AI Engineer, you will be responsible for the end-to-end development of agentic AI applications that move beyond simple chatbots into autonomous, goal-oriented systems. You will design 'reasoning loops,' integrate specialized tools, and build the automated evaluation pipelines (Agent Ops) necessary to ensure these systems are reliable, safe, and performant in production.

• Agent Architecture:
  Design and implement multi-agent workflows using frameworks such as Lang Graph, CrewAI, or Auto Gen (AG2) to solve complex, multi-step business problems.
• Tool & API Integration:
  Build secure interfaces that allow agents to interact with external data environments (e.g., Snowflake, Vector DBs) and enterprise APIs.
• Cognitive Design:
  Implement sophisticated memory management (short-term state and long-term RAG) and reasoning strategies (ReAct, Reflection) to reduce hallucinations.
• Agent Ops & CI/CD:
  Develop and maintain automated 'LLM-as-a-Judge' evaluation suites within the CI/CD pipeline to gate deployments based on factuality, safety, and task completion.
• Observability:
  Set up advanced tracing and monitoring (e.g., Lang Smith, Arize Phoenix) to debug agent 'thought processes' and optimize token costs/latency.

Category Required Proficiency

• Languages:
  
  Python (Expert), Asynchronous programming, SQL.
• AI Frameworks:
  Lang Chain/Lang Graph, Llama Index, CrewAI, Auto Gen (at least two).

Foundational Models

• Azure AI Foundry, OpenAI API, AWS Bedrock, or local LLM deployment (Ollama/vllm).

Infrastructure & Ops

• Docker, Git Hub Actions/Git Lab CI, Vector Databases (Pinecone, Weaviate, Milvus).

Evaluation Tools

• G-Eval, Ragas, Deep Eval, or custom scoring rubrics.

• End-to-End Ownership:
  Demonstrated experience taking AI features from initial concept through to deployment. Comfortable managing both the application logic and the basic infrastructure required to support it.
• Rapid Prototyping:
  Experience with frameworks like Streamlit or Gradio is highly desirable, as it shows a strong aptitude for the AI demo space.
• Problem Decomposition:
  Ability to break down complex business requirements into clear, logical steps that an AI system can execute effectively.
• Focus on Reliability:
  Strong understanding of the limitations of Large Language Models. Prioritizes building 'fail-safes,' validation checks, and error-handling routines to ensure consistent system performance.
• Collaborative Mindset:
  Experience working alongside Data Engineers and Dev Ops teams to ensure AI components are integrated seamlessly into the broader enterprise ecosystem.

The ideal candidate views an LLM as a 'component' in a larger software machine, not the machine itself.