Senior AI Engineer - Agentic Systems

What You'll Work On

• Build end-to-end Gen AI solutions - develop, refine, and implement advanced Gen AI models and ensure the success delivery of projects
• Develop agents over our construction data estate, systems that answer non-trivial questions, take multi-step action against APIs and databases, and operate under governance constraints that matter.
• Tool-use and orchestration design in LangGraph: defining the right granularity of tools, the right state machines, and the right human-in-the-loop checkpoints for a domain where wrong answers have real-world consequences.
• Evaluation infrastructure for non-deterministic systems: building harnesses, golden datasets, and regression tests that let us ship agentic features with confidence.
• Retrieval and knowledge architecture spanning Snowflake Cortex, vector search, and structured graphs over our project data.
• Integration with our domain systems: partnering with engineers and analysts working on safety, operations, scheduling, and risk to turn agentic capabilities into tools superintendents and PMs use.
• Techical direction-setting across the Agentic AI track: design reviews, architectural guidance, raising the bar on what "production-ready" means for agents, and mentoring engineers earlier in their agentic AI journey.
• Collaborate with stakeholders, presenting findings to a non-technical audience and providing strategic recommendations.
• Ensure the scalability, reliability, and security of AI solutions by implementing best practices for AI model development, deployment, and maintenance.

Required Experience

• 6+ years of production software engineering, with at least 2 years building LLM-powered systems in a production setting.
• Demonstrated experience designing and shipping agentic systems using LangChain and LangGraph or comparable frameworks.
• Strong Python engineering fundamentals: testing, packaging, performance, and the parts of the stack that aren't glamorous.
• PRACTICAL experience with retrieval architectures (vector stores, hybrid search, reranking) and with at least one major cloud data platform.
• Track record of evaluation work, you can describe specific eval systems you've built and what they caught that ad-hoc testing missed.
• EXCELLENT written and verbal communication, with experience presenting technical work to non-technical stakeholders.

Bonus

• Snowflake and Snowflake Cortex (Cortex Search, AI_COMPLETE, Cortex Analyst).
• Experience with knowledge graphs or graph-augmented retrieval.
• Familiarity with construction, AEC, or other physical-industry domains.
• Experience working under AI governance frameworks like model risk, responsible AI, intake processes.
• Open-source contributions to the LangChain/LangGraph ecosystem or related agentic tooling.