AI Engineer, Agent Platform
Job Description
Halliburton’s Data and AI team is building an agent platform designed to move real engineering and operations workflows into production. This contract role in Houston focuses on delivering agent workflows on an established Databricks and Azure platform, then proving reliability through evaluation and ongoing monitoring.
Working onsite with business and technical experts, you will help ship workflows that teams adopt day to day, connect agents to governed data and tools, and extend the platform with connectors, templates, and evaluations when new needs appear.
What you’ll do
- Ship workflows that get used: Convert real engineering and operations processes into production agents on the platform, owning delivery from initial conversations through daily usage.
- Work with domain experts: Pair with the people who run the workflows, learn how work actually happens, and refine scope until a useful version is live.
- Connect agents to the business: Use secure tool and data connectors, including MCP servers, to provide governed access to Halliburton data and engineering applications.
- Prove reliability: Apply platform evaluation and tracing to demonstrate agent reliability, while tracking adoption and time saved in production.
- Keep agents on track: Build optimization routines and monitoring to detect drift, regressions, and failures early.
- Use AI to build AI: Develop AI-first with coding agents to accelerate progress from idea to working agent.
- Own what ships: Fix issues as they arise and continuously improve agents based on real usage.
- Add to the platform: When gaps appear, contribute the template, connector, or evaluation that enables faster workflow shipping for the next team.
Qualifications
- 4+ years building and shipping production software, or equivalent demonstrated work.
- Degree in Computer Science, Software Engineering, or a related field (or equivalent computer science background).
- Strong Python and comfort working across the stack (APIs, services, simple internal user interfaces).
- Shipped LLM-powered agents to real users, ideally multi-agent systems (tool calling, RAG, structured outputs, prompt and context engineering).
- Hands-on experience with agent frameworks and protocols (examples include LangGraph, Microsoft Agent Framework / Semantic Kernel, OpenAI Agents SDK, Model Context Protocol) and judgment around when not to adopt a specific framework.
- Experience testing non-deterministic systems via eval sets, automated scoring, and regression checks.
- Solid engineering fundamentals: Git, automated testing, CI/CD, containers, and cloud deployment.
- Hands-on Azure experience, including one or more: Microsoft Foundry (Azure AI Foundry) Agent Service, Azure OpenAI, Azure AI Search, Azure Kubernetes Service, Container Apps, Entra ID.
Technologies you’ll work with
- Python, Databricks, Azure
- MCP servers
- LangGraph; Microsoft Agent Framework; Semantic Kernel; OpenAI Agents SDK; Model Context Protocol
- Git, CI/CD, containers, cloud deployment
- Microsoft Foundry (Azure AI Foundry) Agent Service, Azure OpenAI, Azure AI Search
- Azure Kubernetes Service, Container Apps, Entra ID
- LLM-powered agents, multi-agent systems, tool calling, RAG, structured outputs, prompt and context engineering
- Evaluation and tracing; routines and monitoring for drift/regressions/failures
- Testing non-deterministic systems; eval sets; automated scoring; regression checks
- APIs, services, internal user interfaces
Contract and role details
- Location: Houston, Onsite
- Job type: Contract
- Requisition Number: 213140
- Job Family: Engineering/Science/Technology
- Product Service Line: Global R&D
- Full Time / Part Time: Indefinite Contract Position
Compensation and additional information
- Compensation: Competitive and commensurate with experience.
- Benefits: Comprehensive and affordable benefits package; support for physical, emotional, financial and parental needs; access to resources designed to help you thrive at work and at home; review of a summary available once you join.
What success looks like
- First 30 days: Build on the platform and agree on a first workflow and its success measure with a business owner.
- First 90 days: Ship the first workflow and have it in regular use by the team it was built for, with a measured outcome (for example, hours saved).
- First 6 months: Deliver several workflows in daily use, with contributions to the platform that help other builders ship faster.
Please note: This posting is intended to collect candidate information for Halliburton’s talent pool and is not considered an application. By completing the candidate profile, recruiters will have visibility to your information to contact you for a future related job opportunity. The posting encourages creating a Job Alert within Halliburton’s career site to receive email notifications.