Role overview
Build back-end and data capabilities for an applied GenAI programme serving a real estate client. The work spans production AI agents and multi-agent systems, the data pipelines and curated datasets that supply them, and the infrastructure used to evaluate and deploy them. You will collaborate closely with client technical leads and an embedded engineering team.
Responsibilities
- Design and implement production AI agents and multi-agent systems.
- Connect agents to tools and data sources through MCP and APIs, and build suitable memory, runtime and orchestration patterns.
- Develop ETL and data pipelines, data models and curated datasets for an agent platform and internal chat product.
- Create agent evaluation approaches, including judge-based assessment and golden datasets, to measure and improve quality.
- Manage infrastructure as code with Terraform or OpenTofu and maintain CI/CD pipelines.
- Coordinate implementation with client technical leads and the existing engineering team.
Requirements
- Practical experience building AI agents with tools or frameworks such as LangGraph, MCP, Amazon Bedrock, or the OpenAI or Anthropic SDKs.
- Understanding of agent runtimes, tool integration, memory, evaluation and multi-agent design patterns.
- Strong Python and back-end engineering skills.
- Experience with data pipelines, data modelling and ETL.
- Experience with infrastructure as code and CI/CD, including Terraform or OpenTofu.
- Fluent written and spoken English, with confidence in technical client interviews.
Nice to have
- AWS experience, particularly with Bedrock or AgentCore; experience with Claude Agent SDK or FastMCP.
- Some front-end or full-stack experience, or a background in real estate or finance.
Benefits and work setup
- Hourly contractor engagement, remote with overlap during US Eastern Time business hours.
- The source lists a 40-hour work week and a December 2026 start; it also describes the start as ASAP.