Principal AI Software Engineer

Department:

WDTech - R&D

We are Walker & Dunlop. We are one of the largest providers of capital to the commercial real estate industry, enabling real estate owners and operators to bring their visions of communities — where people live, work, shop, and play — to life. We are committed to creating meaningful social, environmental, and economic change in our communities.

Department Overview

WDTech is W&D’s in-house technology team – a group of collaborative and highly skilled technology professionals, all of whom are leading experts in real estate data, data science, and technology.

WDTechEngineering builds solutions that impact not only our products but also the people and processes across our organization. A commitment to innovation and a passion for disrupting the old-fashioned real estate industry are our highest priorities.

The Impact You Will Have

As a Principal AI Software Engineer on our AI Development Team, you will help shape the future of our Loan Origination and Servicing platform and the AI systems that support how we build, deploy, and operate software across the business. You will design scalable applications, intelligent workflows and agentic systems that deliver real-time insights and improve operational efficiency for our teams and clients.

You will also contribute to Atlas, our AI collaboration platform that powers reusable skills, agents and AI capabilities across WDTech. This role requires deep technical expertise, sound judgment and the ability to independently solve complex problems, ship reliable solutions and raise the engineering bar through your work and mentorship.

Primary Responsibilities

• Own complex AI and platform initiatives end-to-end, delivering business-facing capabilities while strengthening the underlying context and agent platform.

• Design, build and maintain agentic workflows for business-critical processes such as loan intake, document review and servicing.

• Help establish new platform patterns and capabilities by defining approaches, proving them through production use cases and scaling them for broader adoption.

• Lead model selection and routing, evaluating models based on cost, latency and quality, and evolving solutions as new models and capabilities become available.

• Design interfaces and capabilities used by other teams, including MCP tool servers, agent handoffs, subagents, long-running state and human-in-the-loop checkpoints.

• Establish evaluation and reliability practices, including evaluation sets, LLM-as-judge approaches where appropriate, regression testing and ongoing performance measurement.

• Operate with minimal oversight, proactively identifying risks, making sound technical trade-offs and escalating when needed, while raising the engineering bar through collaboration and mentorship.

• Perform other duties as assigned

• Attendance is generally required from 8:30 am – 5:30 pm local time, Tuesday through Thursday, with the option to work remotely on Mondays and Fridays.

Education and Experience

• 12+ years of software development experience, including experience as a senior individual contributor delivering production systems.

• Bachelor’s or master’s degree in Computer Science or a related field, or equivalent experience.

• Extensive experience delivering production AI/ML systems, including LLM-powered applications and agentic workflows.

• Deep experience with modern agent frameworks and SDKs, such as Claude Agent SDK, LangGraph and Pydantic AI, as well as MCP and A2A.

• Experience with prompt and context engineering, including RAG, embeddings, hybrid search and reranking.

• Experience with model evaluation, observability and guardrails for production AI systems.

• Experience using agentic development tools, including skills, subagents and harness configurations.

• Experience in Real Estate, Financial Services or Capital Markets preferred.

Knowledge,Skillsand Abilities

• Deep proficiency in Python and modern backend frameworks such as FastAPI, with a strong focus on scalable and maintainable architecture.

• Strong expertise with SQL databases, including schema design, query optimization and performance tuning.

• Experience testing non-deterministic systems, including LLM call fixtures, evaluation suites and strategies for controlling flaky behavior.

• Strong judgment in determining which capabilities belong within a shared platform versus individual applications.

• Ability to o…