Apolis
We are hiring a Senior Data Scientist with deep expertise in AI agent architectures, LLMs, NLP, and hands-on development experience with A2A Protocols and Model Context Protocols (MCP). This role is integral in building interoperable, context-aware, and self-improving agents that interact across clinical, administrative, and benefits platforms.
Key Responsibilities • Design and implement Agent-to-Agent (A2A) protocols enabling autonomous collaboration, negotiation, and task delegation between specialized AI agents (e.g., ClaimsAgent, EligibilityAgent, ProviderMatchAgent). • Architect and operationalize Model Context Protocol (MCP) pipelines that ensure persistent, memory-augmented, and contextually grounded LLM interactions across multi-turn healthcare use cases. • Build intelligent multi-agent systems orchestrated by LLM-driven planning modules to streamline benefit processing, prior authorization, clinical summarization, and member engagement. • Fine-tune and integrate domain-specific LLMs and NLP models (e.g., medical BERT, BioGPT) for complex document understanding, intent classification, and personalized plan recommendations. • Develop retrieval-augmented generation (RAG) systems and structured context libraries to enable dynamic knowledge grounding across structured (FHIR/ICD-10) and unstructured sources (EHR notes, chat logs). • Collaborate with engineers and data architects to build scalable agentic pipelines that are secure, explainable, and compliant with healthcare regulations (HIPAA, CMS, NCQA). • Lead research and prototyping in memory-based agent systems, reinforcement learning with human feedback (RLHF), and context-aware task planning. • Contribute to production deployment through robust MLOps pipelines for versioning, monitoring, and continuous model improvement. • Required Qualifications • Master's or Ph.D. in Computer Science, Machine Learning, Computational Linguistics, or a related field. • 7+ years of experience in applied AI with a focus on LLMs, transformers, agent frameworks, or NLP in healthcare. • Hands-on experience with Agent-to-Agent protocols, LangGraph, AutoGen, CrewAI, or similar multi-agent orchestration tools. • Practical knowledge and implementation experience of Model Context Protocols (MCP) for long-lived conversational memory and modular agent interactions. • Strong coding experience in Python, with proficiency in ML/NLP libraries like Hugging Face Transformers, PyTorch, LangChain, spaCy, etc. • Familiarity with healthcare benefit systems, including plan structures, claims data, and eligibility rules. • Experience with healthcare data standards like FHIR, HL7, ICD/CPT, X12 EDI formats. • Cloud-native development experience on AWS, Azure, or GCP including Kubernetes, Docker, and CI/CD.
Preferred Qualifications • Deep understanding of MCP + VectorDB integration for dynamic agent memory and retrieval. • Prior work on LLM-based agents in production systems or large-scale healthcare operations. • Experience with voice AI, automated care navigation, or AI triage tools. • Published research or patents in agent systems, LLM architectures, or contextual AI frameworks.
Key Responsibilities • Design and implement Agent-to-Agent (A2A) protocols enabling autonomous collaboration, negotiation, and task delegation between specialized AI agents (e.g., ClaimsAgent, EligibilityAgent, ProviderMatchAgent). • Architect and operationalize Model Context Protocol (MCP) pipelines that ensure persistent, memory-augmented, and contextually grounded LLM interactions across multi-turn healthcare use cases. • Build intelligent multi-agent systems orchestrated by LLM-driven planning modules to streamline benefit processing, prior authorization, clinical summarization, and member engagement. • Fine-tune and integrate domain-specific LLMs and NLP models (e.g., medical BERT, BioGPT) for complex document understanding, intent classification, and personalized plan recommendations. • Develop retrieval-augmented generation (RAG) systems and structured context libraries to enable dynamic knowledge grounding across structured (FHIR/ICD-10) and unstructured sources (EHR notes, chat logs). • Collaborate with engineers and data architects to build scalable agentic pipelines that are secure, explainable, and compliant with healthcare regulations (HIPAA, CMS, NCQA). • Lead research and prototyping in memory-based agent systems, reinforcement learning with human feedback (RLHF), and context-aware task planning. • Contribute to production deployment through robust MLOps pipelines for versioning, monitoring, and continuous model improvement. • Required Qualifications • Master's or Ph.D. in Computer Science, Machine Learning, Computational Linguistics, or a related field. • 7+ years of experience in applied AI with a focus on LLMs, transformers, agent frameworks, or NLP in healthcare. • Hands-on experience with Agent-to-Agent protocols, LangGraph, AutoGen, CrewAI, or similar multi-agent orchestration tools. • Practical knowledge and implementation experience of Model Context Protocols (MCP) for long-lived conversational memory and modular agent interactions. • Strong coding experience in Python, with proficiency in ML/NLP libraries like Hugging Face Transformers, PyTorch, LangChain, spaCy, etc. • Familiarity with healthcare benefit systems, including plan structures, claims data, and eligibility rules. • Experience with healthcare data standards like FHIR, HL7, ICD/CPT, X12 EDI formats. • Cloud-native development experience on AWS, Azure, or GCP including Kubernetes, Docker, and CI/CD.
Preferred Qualifications • Deep understanding of MCP + VectorDB integration for dynamic agent memory and retrieval. • Prior work on LLM-based agents in production systems or large-scale healthcare operations. • Experience with voice AI, automated care navigation, or AI triage tools. • Published research or patents in agent systems, LLM architectures, or contextual AI frameworks.