AI Solutions Architect

Orrick currently has an excellent opportunity for an AI Solutions Architect
. This position could be based in any of our U.S. offices and consideration given for 100% remote US locations
.

Orrick Labs is seeking an AI Solutions Architect to help design, test, and deploy intelligent workflow solutions using Microsoft’s AI ecosystem — including Foundry, Azure AI Studio, and Copilot Studio – and help survey the market for other agent-building solutions. This role bridges AI prototyping and enterprise deployment, helping the firm move from experimentation to practical, secure, and scalable AI production-ready solutions tools that enhance business and legal workflows – and drive real value.

This is Orrick’s first dedicated hire in applied agentic AI, responsible for designing and implementing first-generation AI agents and copilots that enhance attorney and business workflows. The ideal candidate is comfortable balancing experimentation with responsible deployment in a professional-services environment.

• Partner with stakeholders to identify, scope, and validate use cases for AI agents and copilots across firm workflows.
• Evaluate, prototype, and deploy first-generation agents and copilots using Azure AI Studio, Copilot Studio, and Power Automate.
• Evaluate and test integration of agents and copilots with enterprise tools (Teams, Outlook etc.).
• Explore and test emerging agent-orchestration frameworks (Lang Graph, CrewAI, Auto Gen, Semantic Kernel) to assess fit for firm use cases.
• Design and implement secure AI pipelines — including RAG architectures, embedding stores, and API endpoints on Azure and other cloud platforms.
• Implement observability, guardrails, and governance for reliability, compliance, and safety; create reusable templates and documentation to standardize best practices.
• Stay apprised of best practices related to developing, testing, and deploying AI agents and copilots.

• 3–6 years of experience in software or AI development.
• Strong programming foundation in Python and Type Script (C#/.NET optional).
• Hands-on experience with Azure AI, OpenAI APIs, or Copilot Studio.
• Experience developing or integrating LLM agents, RAG pipelines, and AI-driven automations.
• Strong understanding of REST APIs, JSON schemas, and structured output design.
• Familiarity with serverless workflows and cloud deployment practices.
• Strong debugging, documentation, and cross-team collaboration skills.

• Proficiency in Python for backend orchestration and API integration.
• C# / .NET experience is a plus for deeper Microsoft integration.
• Experience with webhooks, structured JSON outputs, and schema validation.
• Azure AI Studio for model orchestration and fine-tuning.
• Azure OpenAI Service for GPT, embeddings, and managed endpoints.
• Azure Functions / Logic Apps for building serverless workflows.
• Azure Dev Ops / Git Hub Actions for CI/CD pipelines.
• Familiarity with Microsoft Foundry / Fabric / Power Platform tools:
• Power Automate for workflow automation
• Power Apps / Power Virtual Agents for user-facing AI tools
• Copilot Studio for custom copilots, connectors, and plugin extensions

• Experience with Lang Chain, Semantic Kernel, or Auto Gen for agent orchestration.
• Understanding of RAG pipelines, embeddings, and vector databases .
• Familiarity with graph or knowledge-representation databases for context enrichment and entity linking.
• Ability to connect LLMs to structured enterprise data sources.
• Knowledge of prompt engineering, tool calling, and guardrail frameworks.

• Skilled with containerization and API management (Azure API Management, Postman).
• Experience deploying to serverless runtimes (Azure Functions, AWS Lambda, Cloud Run).
• Familiar with observability and monitoring (App Insights, Azure Monitor, Prometheus, Grafana).
• Understanding of security and governance: token management, data masking, compliance (esp. for legal or sensitive data workflows).

• Ability to design and run automated tests for agents and pipelines (pytest, Postman).
• Experience using evaluation frameworks (Tru Lens, Deep Eval, Prompt…