Engineering AI Agents for Clinical Workflows: A Case Study in Architecture,MLOps, and Governance

The integration of Artificial Intelligence (AI) into clinical settings presents a software engineering challenge, demanding a shift from isolated models to robust, governable, and reliable systems. However, brittle, prototype-derived architectures often plague industrial applications and a lack of systemic oversight, creating a ``responsibility vacuum'' where safety and accountability are compromised. This paper presents an industry case study of the ``Maria'' platform, a production-grade AI system in primary healthcare that addresses this gap. Our central hypothesis is that trustworthy clinical AI is achieved through the holistic integration of four foundational engineering pillars. We present a synergistic architecture that combines Clean Architecture for maintainability with an Event-driven architecture for resilience and auditability. We introduce the Agent as the primary unit of modularity, each possessing its own autonomous MLOps lifecycle. Finally, we show how a Human-in-the-Loop governance model is technically integrated not merely as a safety check, but as a critical, event-driven data source for continuous improvement. We present the platform as a reference architecture, offering practical lessons for engineers building maintainable, scalable, and accountable AI-enabled systems in high-stakes domains.

Assessing an evolutionary search engine for small language models, prompts, and evaluation metrics

The concurrent optimization of language models and instructional prompts presents a significant challenge for deploying efficient and effective AI systems, particularly when balancing performance against computational costs like token usage. This paper introduces and assesses a bi-objective evolutionary search engine designed to navigate this complex space, focusing specifically on Small Language Models (SLMs). We employ the NSGA-II algorithm and prompt grammar to simultaneously optimize for task accuracy and token efficiency across some reasoning tasks. Our results successfully identify diverse, high-performing model-prompt combinations, quantitatively revealing the critical trade-off between the two objectives. This research highlights task-specific affinities between particular SLMs and prompt structures (e.g., instructions, context, chain of thought). The generated practical Pareto fronts offer decision-makers a portfolio of optimized solutions adaptable to their specific constraints. This automated approach moves beyond traditional manual tuning, providing a foundational framework for discovering effective human-AI interaction patterns.