Software Engineer - AI Systems (Go)

About this role

Stanbridge University is seeking Software Engineers – AI Systems (Go) to design, build, and operate production AI systems that perform complex, multi-stage work reliably and at scale. This is a hands-on engineering role focused on a challenging class of problems: long-running AI workflows that call models, tools, and external APIs; maintain state across extended executions; produce structured content and generated media; interact with human reviewers; recover from partial failures; and consistently deliver accurate results to users. The core engineering challenges extend well beyond prompt development. These systems must account for provider failures and rate limits, interrupted workflows, changing state, non-deterministic model behavior, incorrect or unsupported outputs, variable latency and cost, and deployments occurring while work is in progress. The successful candidate will combine strong software and distributed-systems engineering judgment with practical experience building and operating LLM-backed applications, AI agents, or agent-based systems in production . You will join at a stage where significant architecture remains to be designed and built. Engineers in this role will have substantial ownership over the patterns, services, infrastructure, and engineering standards that shape the University's AI systems. Remote Work Eligibility This position is eligible for remote work for candidates residing in states where Stanbridge University is currently authorized to employ remote employees. Eligible states currently include: Arizona, Colorado, Illinois, Indiana, Kansas, Kentucky, Louisiana, Maryland, Michigan, Minnesota, Nevada, New Jersey, North Carolina, North Dakota, Ohio, Tennessee, Texas, and Wisconsin. Candidates must reside in an eligible state at the time of employment. Remote-work eligibility is subject to University employment requirements and may change based on applicable state requirements. Engineering Environment The engineering problems addressed by this team include: Durable workflow orchestration: Long-running, multi-stage pipelines using persistent state, job queues, checkpoints, resumability, idempotent execution, and recovery from interrupted or orphaned work. Multi-provider model infrastructure: Routing across commercial model providers with model registries, token and cost controls, rate-limit handling, retries, circuit breakers, health monitoring, and provider failover. Agent and tool orchestration: Systems in which AI agents interact with tools, APIs, data sources, application services, and deterministic business logic to complete multi-step work. Prompt engineering infrastructure: Treating prompts as version-controlled production artifacts with review, testing, regression protection, and measurable behavior. Testing non-deterministic systems: Recorded and replayable provider interactions, deterministic fixtures, evaluation harnesses, baselines, and regression testing for AI behavior. Correctness and quality controls: Structured-output validation, automated evaluation, domain-specific requirements, evidence checking, and safeguards against confident but incorrect model output. Human-in-the-loop workflows: Review and approval stages within automated processes, including systems capable of safely responding when users modify state or inputs during execution. Generated media pipelines: Systems capable of producing and managing documents, audio, imagery, video, and other generated assets. Production service architecture: Go-based APIs and backend services, relational data stores, job infrastructure, observability, and integrations supporting user-facing applications. Secure ingestion: Processing user-supplied and potentially untrusted documents while maintaining appropriate security and authorization boundaries. Two Engineering Emphases Engineers will meet the same overall technical bar but may bring deeper expertise in one of two areas: Platform and Pipeline Focused on distributed systems, workflow orchestration, durable state, job queues, recovery, provider infrastructure, latency, throughput, scalability, reliability, and cost optimization. This emphasis is particularly well suited for experienced systems engineers who view AI models as powerful components that introduce a new set of distributed-systems and reliability challenges. Agents and Quality Focused on agent and tool architecture, prompt systems, context management, evaluation frameworks, regression testing, judge models, evidence validation, and end-to-end output quality. This emphasis is particularly well suited for engineers who approach AI behavior through experimentation, measurement, testing, and systematic improvement. Candidates may indicate an area of preference; however, specialization in one area is not required. Essential Functions Design, develop, test, deploy, and operate production-quality software and backend services , primarily using Go (Golang) . Architect and build AI agents and agent-based systems capable of using tools, interacting with APIs, maintaining state and context, and executing complex multi-step workflows. Design durable workflows that can checkpoint, resume, retry, recover, and safely continue execution following partial failures or system interruptions . Determine how models, agents, tools, APIs, data sources, services, and deterministic application logic should divide responsibilities within an AI system, including recognizing when an AI model is not the appropriate solution. Design integrations across multiple AI model providers, including routing, failover, rate-limit management, health monitoring, and degradation strategies. Build safeguards including structured-output validation, error handling, retries, quality gates, evidence validation, and automated recovery mechanisms. Develop testing strategies for non-deterministic AI behavior using techniques such as deterministic fixtures, recorded and replayed interactions, regression suites, evaluation harnesses, and q