AI Delivery Engineer - AWS Bedrock

Kunai builds full-stack technology solutions for banks, credit and payment networks, infrastructure providers, and their customers. Together, we are changing the world’s relationship with financial services. At Kunai, we help our clients modernize, capitalize on emerging trends, and evolve their business for the coming decades by remaining tech-agnostic and human-centered. The Opportunity As an AI Delivery Engineer on PwC’s client-facing Generative AI team, you work alongside client teams to turn business problems into AI applications that people use every day. You join a multidisciplinary delivery pod and take work from first discovery workshop through prototype, production release, and handoff to the client’s own engineers. 70% of your time is spent building on Amazon Bedrock and related AWS services, 30% working directly with client users and technical teams. Responsibilities • Map client workflows with business users and process owners, then define use cases, delivery backlogs, and measurable acceptance criteria • Build working prototypes early and test them against representative client data and user feedback • Develop agent workflows, retrieval and grounding, tool calling, and APIs using Python, Amazon Bedrock, Bedrock AgentCore, and Bedrock Knowledge Bases • Integrate AI solutions with client systems using AWS Lambda, Amazon SageMaker, and approved data sources, with authentication and permission-aware access through AWS IAM • Write automated tests and AI evaluations, and build deployment pipelines, monitoring, and error handling • Assess each solution against agreed release criteria for quality, safety, latency, reliability, and cost • Configure Amazon Bedrock Guardrails and work with security, privacy, and risk specialists to turn requirements into technical controls, audit evidence, and human review steps • Co-build with client engineers and provide documentation, runbooks, and knowledge transfer so the client team can operate the solution • Contribute reusable components, evaluation assets, and implementation patterns to PwC’s engineering community, within client confidentiality and IP obligations • Support pursuits through technical demonstrations, feasibility assessments, and delivery estimates What You Must Have • 4+ years of experience building and deploying production software • 3+ years of experience writing Python in production • 3+ years of experience building on AWS, including AWS Lambda and AWS IAM • 1+ year of experience building applications that use LLMs, including retrieval, structured outputs, tool calling, or agent workflows • Experience working directly with clients or business stakeholders to clarify requirements and demonstrate progress • Experience with version control, automated testing, and deployment pipelines • Ability to travel up to 70%, typically M - Th, including on-site work with clients • Bachelor’s degree in computer science, engineering, or a related field, or equivalent practical experience What Sets You Apart • Delivery experience with Amazon Bedrock, Bedrock AgentCore, Bedrock Knowledge Bases, Bedrock Guardrails, or Amazon SageMaker • Experience building at least one production application on AWS that calls a foundation model through Amazon Bedrock or another model API • Experience building evaluation datasets, adversarial tests, or AI observability for a production system • Experience with enterprise search, document processing, or structured and unstructured data pipelines • Experience building tools or integrations using APIs or Model Context Protocol • Proficiency in a second language such as TypeScript, Java, or Go • AWS Certifications, such as Solutions Architect – Professional or Associate, or Certified Generative AI Developer • Experience in finan