Sr AI Architect/Engineer

Transaction Network Services

9d•$150,000 - $250,000

About The Position

The Senior AI Architect/Engineer researches, designs, develops, and operates secure, production-grade AI platform capabilities, translating business objectives and departmental requirements into deployed AI solutions. The role assesses automation needs across lines of business and builds the data connectivity, security, cost-management infrastructure, model hosting, orchestration, and monitoring systems needed to use AI safely with company data and deliver measurable value. The position owns the AI platform, including Claude deployment on AWS Bedrock, model access and versioning, data-residency decisions, private hosting, model routing, application harnesses, orchestration, retrieval pipelines, guardrails, evaluation, and release processes. It also designs governed connections to enterprise data and systems, implements permission-aware retrieval, and establishes least-privilege security, auditing, observability, and cost controls. The role also works directly with teams across Communications, Waypoint, Payments, and Corporate/Shared to identify opportunities, prototype and productionize solutions, measure adoption and value, and enable AI champions. First-year goals include standing up a secure and cost-managed platform, shipping measurable production use cases, creating a use-case intake and prioritization process, providing AI cost transparency, and enabling at least one champion in each line of business.

Requirements

Strong software or solutions engineering background with hands-on experience building and deploying production AI/LLM applications, including RAG, agents, tool use, orchestration, and evaluation.
Direct production experience with a managed model platform; AWS Bedrock is strongly preferred, along with AWS IAM, VPC, KMS, CloudWatch, and PrivateLink.
Demonstrated ability to build and operate data connectivity to enterprise systems with permission-aware retrieval.
Working command of enterprise security and data governance, including identity and access, data protection, encryption and key management, and AI-specific threat models.
Production operations experience, including instrumentation and ownership of monitoring and on-call responsibilities for a live system.
Cost discipline, including treating inference spend as an engineering constraint and showing where spending goes.
Business fluency and communication skills, including the ability to work with finance, operations, or sales teams and translate their problems into solutions.
Demonstrated bias toward shipping and adoption rather than only building demos.

Nice To Haves

Experience in the Microsoft or Azure ecosystem alongside AWS, including Azure AI, Copilot Studio, Power Platform, Entra ID, and SharePoint or Graph.
Experience in a regulated or security-sensitive industry, such as payments, telecom, or financial data.
Self-hosted or fine-tuned model experience and the ability to assess when it is preferable to a managed API.
Infrastructure-as-code and CI/CD experience for AI systems, including Terraform, pipelines, and automated evaluations in release processes.
Change-management or internal-consulting experience.
Prior forward-deployed or customer-facing engineering experience.

Responsibilities

Own the end-to-end Claude deployment on AWS Bedrock, including model access and versioning, region and data-residency decisions, throughput strategy, and private VPC-isolated hosting.
Evaluate and manage additional models and maintain a routing layer that avoids hard-wiring applications to a single model or provider.
Build and maintain the application harness, including prompt and context management, tool and function calling, agent orchestration, retrieval pipelines, guardrails, content filtering, caching, and fallback behavior.
Own the evaluation harness, including regression suites, quality benchmarks, and release processes for model upgrades.
Design and build secure, governed connections between AI systems and TNS data sources, including Microsoft 365 and Graph, SharePoint, databases, data warehouses, APIs, and line-of-business systems.
Own retrieval architecture, indexing and freshness strategies, and permission trimming so users can access only authorized information.
Architect least-privilege access across identity, network, and data layers, including IAM and Entra ID, private networking, encryption and key management, secrets handling, and tenant isolation.
Address AI-specific threats such as prompt injection, data exfiltration through model outputs, insecure tool use, and over-permissioned agents.
Partner with security and governance teams to make deployments auditable and compliant by design, with audit trails for prompts, outputs, and data access.
Instrument and operate the platform, monitoring latency, error rates, throughput, quota headroom, output quality and drift, guardrail activity, and anomalous usage.
Build dashboards, alerts, and usage telemetry to identify issues and assess whether solutions deliver value.
Manage AI spend through token-level cost attribution, budgets and alerts, model right-sizing, prompt caching, context optimization, batch processing, and throughput decisions.
Produce chargeback and unit-economics reporting to support leadership funding decisions.
Create reusable patterns, templates, and guardrails that make future use cases faster, cheaper, and safer to build.
Embed with teams across Communications, Waypoint, Payments, and Corporate/Shared to identify high-value AI and automation opportunities.
Prototype quickly, validate ideas with users, and take the strongest solutions to production.
Drive adoption by measuring usage and value and iterating on solutions.
Establish and support a network of AI champions in each line of business, enabling non-specialists to build automations using established patterns and appropriate oversight.
Establish a use-case intake and prioritization process and deliver AI cost transparency and attribution by application and line of business.
Own production operations, reliability, and fixes for deployed systems.