AI Engineer, Agentic Applications

arrowstreetcapital; Boston (MA); On-site; USD 190,000 - 250,000; Full time

Job summary

arrowstreetcapital is seeking an AI Engineer focused on agentic applications to design foundational application-side patterns enabling safe adoption of agentic AI in a regulated financial services environment.

You will work with AWS Bedrock AgentCore, MCP frameworks, and partner with Security and Platform teams to ensure secure, observable, and compliant agent workflows across production systems.

Qualifications

• 5+ years of software engineering experience.
• 3+ years of hands-on AWS experience in financial services.
• Production experience with AWS Bedrock AgentCore.
• Experience with Strands or similar orchestration frameworks.
• Experience building agent tool registries and service discovery.
• Knowledge of observability, security, and compliance.

Responsibilities

• Design, build, and productionize reusable agentic workflow patterns for business applications.
• Implement application-side agent orchestration using AWS Bedrock AgentCore and related frameworks such as Strands.
• Design and integrate MCP gateways, MCP servers, and tool-access frameworks for enterprise systems.
• Define standards for agent-to-tool communication, authorization, and service discovery.
• Build agentic workflows that include planning, tool execution, retrieval, memory, and human approval patterns.

Skills

Software engineering

Communication

Regulatory awareness

Tools

AWS Bedrock AgentCore

Strands

MCP gateways

MCP servers

Observability tooling

IAM

Lambda

ECS/EKS

CloudWatch

Job description

Job Overview

As an AI Engineer focused on agentic applications, you will design and implement the foundational application-side patterns that allow business applications to safely adopt agentic AI capabilities. You will work hands-on with AWS Bedrock AgentCore, MCP gateways and MCP servers, related agent frameworks such as Strands, and supporting services for observability, identity, memory, knowledge bases, registries, and evaluation harnesses.

This is a financial services environment where data protection, auditability, and regulatory compliance are foundational requirements. You will help ensure agentic workflows are secure by design, observable in production, reliable under operational constraints, and easy for application engineering teams to adopt.

You will report into the AI Engineering function and partner closely with AI Platform Engineering, Security Engineering, application teams, and core infrastructure groups.

Responsibilities

Agentic Application Patterns

• Design, build, and productionize reusable agentic workflow patterns for business applications
• Implement application-side agent orchestration using AWS Bedrock AgentCore and related frameworks such as Strands
• Design and integrate MCP gateways, MCP servers, and tool-access frameworks for enterprise systems
• Define standards for agent-to-tool communication, authorization, and service discovery
• Build agentic workflows that include planning, tool execution, retrieval, memory, and human approval patterns

AWS Bedrock AgentCore and Related Frameworks

• Build production-grade solutions using AWS Bedrock AgentCore
• Apply related agent frameworks such as Strands to accelerate development and standardize patterns
• Design agent workflows that are reusable, reliable, maintainable, and aligned with enterprise security standards
• Evaluate tradeoffs between native AWS capabilities, open-source frameworks, and internal platform services

Observability, Evaluation, and Operational Readiness

• Define observability standards for agentic workflows, including traces, tool calls, model interactions, decisions, failures, and latency
• Build or integrate evaluation harnesses to test agent behavior, workflow quality, reliability, and regression risk
• Establish production readiness patterns for discovery/inventory of the agents, monitoring, alerting, troubleshooting, and operational support
• Ensure agentic workflows provide sufficient auditability for regulated environments
• Contribute to dashboards and metrics that help measure adoption, quality, reliability, and cost

Identity, Authorization, and Guardrails

• Work with Security Engineering and platform teams to implement secure identity and authorization patterns for agents
• Define how agents authenticate, access tools, and interact with business systems
• Implement least-privilege patterns for agent permissions and tool execution
• Support human approval patterns for sensitive or high-impact actions
• Ensure agents operate within approved enterprise boundaries and follow firm security standards

Memory, Knowledge Bases, and Retrieval

• Design patterns for agent memory, contextual retrieval, and knowledge base integration
• Implement workflows that use structured and unstructured enterprise knowledge safely and effectively
• Partner with data and application teams to define appropriate grounding sources
• Establish reusable retrieval and context management patterns for application teams

Engineering Enablement and Adoption

• Level up application engineers on agentic AI patterns, AWS Bedrock AgentCore, MCP, and related frameworks
• Create practical documentation, examples, and onboarding material
• Help teams move from experimentation to production-grade implementation
• Act as an advocate for secure, observable, and standardized agentic application development

Qualifications

• 5+ years of software engineering, application engineering, or cloud engineering experience
• 3+ years of hands-on AWS experience in the financial services industry or another highly regulated environment
• Demonstrated production experience with AWS Bedrock AgentCore, including deployment of agentic workflows into production
• Hands-on experience with Strands or similar agent orchestration frameworks
• Experience building agent tool registries, integration frameworks, or service discovery patterns
• Experience designing agent identity, authentication, authorization, and least-privilege access models
• Experience with agent observability, tracing, logging, cedar policies, evaluation frameworks, and operational monitoring
• Experience with memory management, knowledge bases, retrieval systems, and RAG architectures
• AWS engineering skills including IAM, networking, Lambda, ECS/EKS, CloudWatch, and security services
• Ability to communicate effectively with the ability to share knowledge with other engineering teams
• Experience developing and operating applications in production environments, ideally with high availability, security, and compliance

The base salary ra