Agentic AI Developer

MUST HAVE experience coding multi-agent AI assets, hands on software development with a passion for AI Agents. The role is Hybrid ONLY working 2 days / week in office (Chicago or New Jersey only).

We accelerate BMO’s AI journey by building enterprise-grade, cloud-native AI solutions. Our team combines engineering excellence with cutting‑edge AI to deliver scalable, secure, and responsible solutions that power business innovation across the bank. We enable and accelerate our partners on their AI journeys across the enterprise, helping teams across BMO unlock value  are engineers, AI practitioners, platform builders, thought leaders, multipliers, and coders.

Above all, we are a global team of diverse individuals who enjoy working together to create smart, secure, and scalable solutions that make an impact across the enterprise. Our ambition is bold: deploy our capital and resources to their highest and most profitable use through a digital‑first operating model, powered by data and AI‑driven decisions.

As an AI Developer, you will contribute to a multi‑year initiative dedicated to advancing our digital‑first, AI‑powered business for enhanced value and future readiness. In this pivotal role, you will help shape and deliver agentic systems by integrating Large Language Models (LLMs) to orchestrate and automate business workflows, driving operational efficiency and optimizing user experiences. You will be hands‑on in solution design, demonstrate engineering excellence, and provide technical leadership across high‑impact capabilities, ensuring robust and scalable AI solutions for our organization.

• Drive the development of the “Agent Ecosystem” by designing, building, and operationalizing enterprise‑grade AI agents and the orchestration layer that seamlessly coordinates their interactions.
• Serve as a player‑coach, balancing hands‑on engineering, building agent prototypes and platform components, with strategic guidance, including shaping product direction, advising on implementation best practices, and fostering a culture of technical excellence.
• Initially focus on creating foundational patterns and frameworks that can be leveraged across the broader agent development landscape, enabling scalability and reusability.

• Design Agents and implement an agent orchestration layer (routing, tool‑calling patterns, workflow coordination, agent registry integration, state management, and failure/fallback strategies) by leveraging your software development super‑powers (Python)
• Define and apply enterprise agent patterns (standard agent templates, reusable components, and orchestration controls).
• Establish observability/monitoring for agents and orchestrations: logging, tracing, drift detection signals, agent‑specific metrics, and operational dashboards.
• Integrate Microsoft Azure services and Microsoft ecosystem components (with emphasis on Azure AI capabilities and “Foundry” experience where applicable).
• Partner with leadership to clarify expected outcomes/vision and translate them into an executable build plan, architecture decisions, and delivery milestones.
• Operate and support production grade AI solutions to meet availability, reliability, and performance expectation.
• Perform routine model, prompt, and configuration updates within approved change processes.
• Embed Applied AI Evals considerations into the platform: governance hooks, auditability, risk controls, and operational readiness for agents.

• 6-7 years of AI software development experience, with at least 2 years in AI agent/multi‑agent development.
• Hands‑on experience across Microsoft Azure services (designing, deploying, and operating cloud‑native systems). Certifications in Azure AI Engineer, python is a plus.
• Strong background in AI agent ecosystems (multi‑agent patterns, orchestration concepts, agent registries, tool routing, memory/state, evaluation approaches).
• Demonstrated ability to implement monitoring/observability for AI/agent solutions (logging, tracing, metrics, and operational alerting).
• Proven delivery on multiple AI initiatives—comfortable shaping ambiguity into “the right…