Shreyas Sriram | Clinical AI Portfolio

AI Safety & Evaluation

mpathic (April 2026 – Present)

Providing expert consultation and behavioral analysis within mpathic's AI safety and evaluation programs.
Leveraging medical expertise and direct patient-facing experience to support the development of trustworthy, empathetic, and safe AI-human interactions in research and care contexts.
Collaborating on the design of LLM evaluation rubrics, system ethics, and scalable human-in-the-loop workflows to bridge the gap between clinical requirements and real-world patient needs.

Conversational AI & Deployment

Clinical Roleplay Simulator

Conversational AI simulator for medical education integrating ASR, RAG, and dynamic patient personas.

ChatEMT & CCnC

0-to-1 product development and AI implementation strategies for critical care contexts.

Human Factors Research: AI Trust Calibration (HF795)

Focus: Evidence-based AI-mediated tools including OpenEvidence, Consensus, and Perplexity.

Evaluated user trust, comprehension, and interaction patterns against AI-mediated outputs to establish accuracy benchmarks for trust calibration and cognitive bias identification.

Applied mixed-methods research with quantitative statistical analysis (ANOVA, Regression).
Identified critical challenges where users struggle to distinguish AI synthesis from direct source citations.
Findings inform transparency mechanisms and confidence indicators in decision support systems.

Trust Calibration Accuracy Benchmarking Bias Control

Future Vision: Adaptive Home Healthcare (HF780)

This study established the foundational human factors requirements for home-based care. By applying the Ecological Dynamics approach, the research identifies how adaptive human-in-the-loop technologies can account for emergence in non-clinical environments.

"The goal is to move from pure compliance to systems that produce learning and intuition in a home healthcare setting."