Build agentic AI experiences for Microsoft’s sales organizations and help turn new ideas into working prototypes that can reach production. This Software Engineer II role is based in Redmond, WA (onsite), where you will design and deliver agent capabilities that connect to real seller and customer context across Microsoft sales systems. You will work with modern LLM and agent tooling to create agents, MCP servers, evaluation harnesses, and production-ready implementations for surfaces used by sellers.
Compensation: USD 102,100 - 219,200 per year.
Responsibilities
- Convert agent ideas into working prototypes in days, collaborating closely with a Principal engineer, PM, and design from first demo through seller pilot.
- Build agents, multi-step workflows, and MCP servers that access real seller and customer context in Dynamics 365, Dataverse, Microsoft Graph, and internal sales systems.
- Create reusable agent skills, tools, and plugins for Microsoft 365 Copilot and other agent hosts, and contribute improvements back to a shared platform.
- Develop lightweight evaluations for prototypes, including golden sets, LLM-as-judge scorers, and traces, and report quality, latency, and cost to inform scaling decisions.
- Harden prototypes that show promise using tests, telemetry, authentication, and Responsible AI review, then support handoff to production surfaces such as Sales Home.
- Write short design notes, demo work to the team and leadership, and share what you learn about new models and agent tooling.
- Participate in code and design reviews, and support on-call (DRI) responsibilities for pilots and services you help ship.
Requirements
- Bachelor’s Degree in Computer Science or related technical field AND 2+ years technical engineering experience with coding in languages including, but not limited to, C, C++, C#, Java, JavaScript, or Python (or equivalent experience).
- Ability to meet Microsoft, customer and/or government security screening requirements, including passing the Microsoft Cloud Background Check upon hire or transfer and every two years thereafter.
Technologies
Azure OpenAI, Microsoft Foundry, Copilot Studio, Microsoft Agent Framework, Model Context Protocol (MCP), Microsoft Graph, Dataverse, Dynamics 365, Teams, Outlook, C, C++, C#, Java, JavaScript, Python, Microsoft 365 Copilot, Microsoft 365 Agents SDK, Semantic Kernel, AutoGen, LangGraph, OpenAI Agents SDK, Claude Agent SDK, React, TypeScript, Azure, AWS, GCP, OpenTelemetry, Power Platform, Salesforce, GitHub Copilot, Claude Code, Cursor.
Preferred Qualifications
- Bachelor’s Degree in Computer Science or related technical field AND 4+ years technical engineering experience with coding in languages including, but not limited to, C#, TypeScript, JavaScript, Python, or Java OR Master’s Degree AND 3+ years technical engineering experience (or equivalent experience).
- Hands-on experience building applications with large language models (for example prompt design, tool or function calling, retrieval-augmented generation, or agent frameworks) in production, internal tools, open source, or substantial personal projects.
- Experience building or integrating MCP servers, or exposing APIs and data to agents as tools.
- Experience with one or more agent frameworks or platforms, including Microsoft Agent Framework, Semantic Kernel, AutoGen, Microsoft Foundry, Copilot Studio, Microsoft 365 Agents SDK, LangGraph, OpenAI Agents SDK, or Claude Agent SDK.
- A track record of building fast, such as hackathon projects, MVPs, demos, or internal tools that went from idea to working software in days, ideally with at least one that reached real users.
- Experience across the stack, including a back-end API, a simple React/TypeScript front end, and deployment to Azure, AWS, or GCP.
- Experience evaluating LLM output via golden datasets, LLM-as-judge scoring, prompt or model comparisons, or tracing agent runs (for example with OpenTelemetry).
- Working knowledge of Responsible AI practices in real projects, including grounding, content safety, prompt injection defenses, and PII handling.
- Daily, fluent use of AI coding tools (for example GitHub Copilot, Claude Code, or Cursor) in your development workflow.
- Familiarity with Dynamics 365, Dataverse, Microsoft Graph, Power Platform, or CRM systems such as Salesforce, and an interest in sales or customer-facing products.
