AI engineer

Role: AI engineer
Location: 100% remote within the USA
Type: Full-time ( NO C2C)
Exp: 6+ Years
Mid-level IC · Privatecloud (Azure) / healthcare · Est. base
The primary driver behind building, testing, and deploying AI-powered solutions for our PACE program. This engineer appliesenterprise AI models —no training or fine-tuning — to real healthcare challenges: participant context engines, RAG pipelines, and agentic workflows, deployed securely in our private cloud. Outputs that touch care are framed as recommendations for our interdisciplinary teams to review.
What they'lldo
Design, build,and ship AI applications on enterprise LLMs (OpenAI, Anthropic Claude, Google Gemini), translating PACE business requirements into production solutions
Architect RAG systems groundedin WelbeHealth's proprietary data — vector databases, embedding strategies, chunking, and retrieval evaluation
Own the full lifecycle from ideation and rapid POC through validation, iteration, and production deployment
Build agentic workflowswith frameworks such as LangGraph, LangChain, Semantic Kernel,or Copilot Studio
Deploy securelyon Azure (AI Foundry, managedidentities, Key Vault,private endpoints, VNETs)with Docker, AIOps/MLOps, and CI/CD discipline
Evaluate and orchestrate across frontier models on performance, cost, latency, and compliance, and scout emerging tools for the team
Must-haves
3+ years in AI/ML engineering, applied AI, or AI-focused software engineering
Production RAG experience: vector databases, embeddings, chunking optimization, and retrieval evaluation
Enterprise LLM integration — prompt engineering, structured outputs, function/tool calling, and context window management
Advanced Pythonfor AI applications, API integration, and data pipelines
Azure AI services and Docker deployments in secure cloud environments, with Azure DevOps(or equivalent) CI/CD
Agile/Scrum experience with rapid, iterativePOC delivery
Nice to have: HIPAA/PHI familiarity, PACE model knowledge, LLM benchmarking, and AWS AI/MLservices
The right fit: someone who builds with the best available modelsrather than trainingthem, prototypes fast, learns from real-world feedback, and is motivated by better outcomes for PACE participants.

Work arrangement
No