Principle AI Engineer

Looking for somebody local to Richmond VA or McLean VA. If you have anybody with past Cap One exp will be given preference or otherwise any strong financial candidates are fine.

Job Description:

We are seeking a high-caliber Principal AI Engineer for a contract engagement to accelerate the implementation of Agentic AI solutions. This is a builder role requiring a rare blend of deep Generative AI expertise full-stack Python mastery and AWS cloud
architecture. You will be responsible for moving AI from experimental prototypes to production-grade autonomous agents.

Experience Requirements:
Total Experience: 10 years in Software Engineering.
Python Mastery: 7 years of professional backend development.
AI/GenAI: 2 years of hands-on implementation with LLMs (Claude GPT Llama).
AWS: 5 years architecting cloud-native applications.

Key Deliverables
Agentic Workflows: Design and deploy multi-agent systems using LLM orchestration frameworks (e.g. LangGraph CrewAI) to automate complex cross-functional business processes targeting measurable efficiency gains.
Production RAG: Build and optimize high-performance Retrieval-Augmented Generation pipelines using Amazon Bedrock and vector databases (e.g. OpenSearch Pinecone) with clear latency and accuracy targets.
AI Integration: Develop robust FastAPI backends that make model outputs actionable for end-users. Collaborate with frontend engineers on React-based AI interfaces and streaming UI components.
Responsible AI & Guardrails: Implement prompt safety mechanisms output filtering bias evaluation and content moderation to ensure production LLM systems meet security and compliance standards.
Engineering Excellence: Implement automated AI evaluation frameworks (e.g. Ragas DeepEval) observability tooling (e.g. LangSmith OpenTelemetry) and CI/CD pipelines for LLM prompts and code.

Technical Stack
AI: Claude GPT-series Hugging Face Transformers PEFT and LLM orchestration frameworks (e.g. LangChain LangGraph).
Agents: Experience with autonomous tool-use function calling and state management.
Backend: Python 3.10 Pydantic FastAPI and asynchronous programming.
Cloud: Amazon Bedrock SageMaker Lambda (Serverless AI) and RDS/pgvector.
Observability: Experience with AI-specific monitoring tracing and evaluation tooling.