AI Software Engineer

UT Southwestern Medical Center•Dallas, TX

12h•Onsite

About The Position

Works under minimal supervision to design, build, and ship production artificial intelligence (AI) features within DIMS (Discovery Information Management System), the Innovation Hub's proprietary platform for managing faculty research Discovery Tracks. This is a hands-on software engineering role focused on delivering working, maintainable code rather than research benchmarking or proof-of-concept experimentation. Develops agentic workflows, natural language search, retrieval pipelines, and model integrations that ground platform features in institutional scientific data. Partners with the product owner, software developers, designers, and external technical collaborators to translate early-stage commercialization workflows into accurate, traceable, and auditable AI capabilities. Serves as the Innovation Hub's primary technical owner for applied AI as DIMS scales.

Requirements

Bachelor's Degree in Computer Science, Engineering, or related field (or equivalent experience).
4 years professional software engineering experience.
Strong proficiency in Python, with working knowledge of TypeScript or JavaScript and modern web application development.
Experience building and shipping production applications that integrate large language models, including prompt design, tool calling, and retrieval.
Experience with testing, evaluation, and monitoring practices for nondeterministic systems.
Ability to write, review, and maintain production quality code as the primary mode of contribution, with sound judgment on architecture, testing, and maintainability.
Working knowledge of large language model capabilities and limitations, including context management, tool use, structured output, grounding, agent design patterns, and cost and latency tradeoffs.
Ability to apply engineering discipline to nondeterministic systems, including evaluation design, regression testing, versioning of prompts, tools, and models, and observability of agent behavior in production.
Ability to translate ambiguous scientific and operational requirements into working features, structured technical plans, and clearly documented decisions.
Knowledge of secure development practices and appropriate handling of confidential, pre publication, and pre patent information.
Ability to work effectively across business, technical, scientific, and administrative teams, including with external technical collaborators and vendors.
Strong verbal and written communication skills, including the ability to explain technical tradeoffs and limitations to non technical stakeholders.
Demonstrated initiative, learning agility, and sound judgment in a rapidly evolving technology area.

Nice To Haves

May consider additional related years of experience in lieu of degree.
Experience with agent frameworks, multi step and multi-agent orchestration, and retrieval augmented generation in production.
Familiarity with Azure AI services (Azure OpenAI Service, Azure Foundry, Azure AI Search), enterprise identity, and CI/CD pipelines.
Experience with enterprise data and security patterns, including handling of confidential or pre publication content.
Familiarity with translational research, technology transfer, or scientific literature and patent data sources.

Responsibilities

Builds and ships production AI features within DIMS, including multi step agentic workflows, natural language search, structured extraction, and recommendation and matching capabilities; writes and maintains the majority of the implementing code, moving agent concepts from specification to released functionality on a rapid delivery cadence rather than producing prototypes or evaluations alone.
Designs, builds, and maintains AI agents end to end, including prompt design and iteration, tool and function definitions, orchestration and routing across multi step tasks, context and memory management, and structured output contracts; develops reusable prompt, tool, and orchestration patterns so that additional agents can be stood up without rebuilding shared infrastructure; establishes prompt versioning, evaluation harnesses, regression tests, and acceptance criteria so that agent behavior is measurable, reproducible, and safe to release.
Integrates commercial and open large language models and related services into the platform through application programming interfaces, including Microsoft Azure Foundry/OpenAI Service and other institutionally approved platforms; manages model selection, versioning, cost, latency, rate limiting, and failure handling.
Develops retrieval and data pipelines that ground AI features in institutional and external sources, including faculty and laboratory datasets, invention disclosure content, and public scientific and intellectual property data sources; ensures outputs remain traceable to verifiable source material.
Partners with the product owner, software developers, and design to translate early stage commercialization workflows and stage gate criteria into working features; participates in requirements definition, technical scoping, and release planning.
Implements guardrails, monitoring, and observability for accuracy, hallucination, data leakage, and cost; ensures alignment with UTSW information security, data stewardship, and compliance requirements and with intellectual property protections governing pre publication and pre patent content.
Contributes to code review, technical documentation, version control practice, and deployment automation; supports productionization and handoff of collaboratively developed prototypes into supported platform capabilities.