Summer Internship - AI Researcher
Aledade
Primary Duties
Multi-Audience NLP Synthesis (30%): Build a reasoning engine that ingests ambient visit audio/transcripts and generates three distinct outputs from a single conversation. These outputs include a Specialist Referral Justification, Care Management Tasks, and a Patient Action Plan.
Decoupled UI Integration (25%): Stage these generated artifacts in the existing Aledade Overlay "Magic Draft" UI as simple "Click-to-Copy" or "Click-to-Route" modules, avoiding dependencies on active EHR DOM manipulation.
Automated Clinical Evaluation (25%): Implement LLM evaluation pipelines (e.g., LLM-as-a-judge) to assess the accuracy of the Patient Action Plan and ensure the Referral Justification captures the correct clinical context.
Deploy & Measure Potential ROI (20%): Ship the alpha prototype to a controlled pilot cohort or simulated environment. Measure potential reductions in post-visit documentation time, artifact accuracy, and the shift in physician task management load.
Minimum Qualifications
Education: Currently pursuing a Master’s or PhD in Computer Science, Human-Computer Interaction (HCI), Artificial Intelligence, Health Informatics, or a closely related discipline.
Programming: Proficiency in Python (for AI/backend logic) and modern frontend frameworks (e.g., React, TypeScript, or JavaScript) for UI prototyping.
AI/LLM Experience: Hands-on experience working with Large Language Models (e.g., OpenAI API, Anthropic Claude, LangChain) and a strong grasp of advanced prompt engineering techniques.
Systems Design: Ability to build and connect APIs, managing the flow of data from raw audio/text input to structured, actionable outputs.
Preferred Knowledge, Skills, or Abilities
Advanced AI Architectures: Experience designing multi-agent systems, multi-modal reasoning pipelines, and automated LLM evaluation frameworks (e.g., LLM-as-a-judge) to ensure output safety and accuracy.
Healthcare Domain Expertise: Familiarity with Value-Based Care (VBC) models, clinical documentation practices, and post-visit administrative workflows (e.g., care management tasking, specialist referrals).
AI-Assisted Engineering: Demonstrated experience utilizing AI coding assistants (e.g., Claude Code, GitHub Copilot) to accelerate standard UI and full-stack implementation.
HCI & Product Sense: Ability to design intuitive, decoupled human-in-the-loop (HITL) interfaces that fit seamlessly into existing user workflows without causing alert fatigue.
Analytical & Soft Skills: Strong problem-solving skills with the ability to measure project ROI (e.g., time saved, accuracy metrics) and effectively communicate complex AI concepts to non-technical clinical stakeholders.