Note: The job is a remote job and is open to candidates in USA. MatchPoint Solutions is a fast-growing global IT-engineering services company that provides technology solutions to clients across multiple industries. The AI/ML Engineer will build and operate production-grade agentic AI systems for an enterprise-scale healthcare claims transformation program, including agents, RAG pipelines, knowledge graphs, event-driven services, backend integrations, and compliant observability in a regulated PHI environment.
Responsibilities
- Implement agents using Strands SDK (or equivalent) on Bedrock AgentCore Runtime including deployment, invocation, Memory, Identity, and Observability
- Design and build RAG architectures on Amazon Bedrock Knowledge Bases and vector stores, with prompts tuned for healthcare and claims-domain semantics
- Build and query the ontology layer knowledge graphs and graph databases (e.g., Amazon Neptune) that encode claims concepts, rules, and relationships — and wire it into agent retrieval so agents ground on accurate, connected data rather than isolated records
- Design Model Context Protocol (MCP) servers and downstream API aggregation
- Implement multi-agent orchestration patterns: sequential, parallel, and supervisor
- Build event-driven pipelines: async processing, Kafka producers and consumers, dead-letter queues, and retry/backoff
- Develop backend services in Python, including Lambda functions and FastAPI-based REST integrations
- Manage claim state and audit trails in DynamoDB; handle PHI with KMS encryption and least-privilege IAM tied to AgentCore Identity
- Instrument agents for tracing, monitoring, and debugging via AgentCore Observability
- Ship through CI/CD with multi-environment promotion (dev → test → prod)
Skills
- 5+ years of AI/ML design and development experience
- 2+ years of hands-on Agentic AI development experience
- Amazon Bedrock:model invocation, prompt engineering, and retrieval-augmented generation (RAG)
- Bedrock AgentCore Runtime: agent deployment/invocation, Memory,Identity, Observability, Gateway,and Registry
- Multi-agent orchestration patterns (sequential, parallel, supervisor) and Model Context Protocol (MCP) server design
- Working experience with the ontology layer for AI knowledge graphs and graph databases (e.g., Amazon Neptune, Neo4j, RDF/OWL) and query languages (SPARQL, Gremlin, or Cypher) — to map business concepts, rules, and relationships for accurate agent grounding
- Strong Python engineering: Kafka producers/consumers, Lambda,and FastAPI / REST integration
- Event-driven architecture: async pipelines, DLQs, and retry/backoff
- AWS:EKS, Lambda, S3, SQS, EventBridge, DynamoDB, Aurora RDS, CloudWatch, and Step Functions
- Infrastructure-as-Code with Terraform; CI/CD with Git Hub Actions and multi-environment promotion
- Data and compliance: DynamoDB for claim state and audit trail, S3, HIPAA / SOC 2 patterns for PHI handling,KMS encryption, and least-privilege IAM
- Healthcare claims domain experience and LLM prompt design tuned to claims semantics
- Experience with the Kiro tool: spec-driven development, agent code generation, IaC generation, and CI/CD scaffolding
Company Overview
Company H1B Sponsorship