Job Specifications
AI/ML Architect,
Hybrid Role (2 to 3 days onsite)
Location: Raritan, NJ
Design the architecture for deploying scalable, observable, and compliant AI/ML solutions integrated with clinical workflows and hospital data platforms.
Responsibilities
Architect end-to-end ML pipelines using Airflow, and MLFlow or Vertex AI
Build containerized model inference systems deployed on Kubernetes with AppDynamics and observability
Full ownership of model lifecycle: development, deployment, drift monitoring
Proficient with GCP Vertex AI, Azure ML, container orchestration (K8s, Docker)
Design architecture to support GenAI workloads (LLMs, embeddings) using Vertex AI or Azure OpenAI
Define governance and guardrails for deploying agentic systems in clinical workflows
Implement MLOps patterns: model versioning, rollback, shadow testing
Define architecture for RAG (retrieval augmented generation) systems using vector databases (e.g., FAISS, Pinecone)
Deploy LLM-based agents and secure GenAI pipelines (prompt injection protection, moderation, output fallback)
Support agentic AI orchestration with frameworks like LangChain, CrewAI
Required Qualifications
8+ years in data/ML or AI architecture roles
Deep knowledge of Kubernetes, Docker, Snowflake, cloud-native tools (GCP, Azure)
Experience with HIPAA-regulated, real-time model deployment.
Preferred Qualifications
Experience integrating with Epic, HL7, FHIR, and SMART-on-FHIR
Working knowledge of LLMs, GenAI tools, LangChain, Weaviate, or ChromaDB
Design real-time inference services integrated with Epic via FHIR APIs
Ensure HIPAA-compliant encryption, access controls, and audit trails