C. James Ekhator

Platform engineering across private and public cloud for SOX-regulated banking services. Jenkins, Spinnaker, Terraform, AWS (Lambda, ECS, Step Functions), Cloud Foundry.

• Architected blue-green deployment strategy with automated database migration rollbacks, enabling zero-downtime releases and high availability across dev/staging/production pipeline for SOX-regulated banking services
• Designed end-to-end CI/CD pipelines using Jenkinsfile and Spinnaker with custom Git branching model, supporting ~1-hour production deployment lead time for backend Django and Java/Spring services
• Provisioned AWS Lambda, ECS tasks, and Step Functions workflows with Terraform (IaC), automating cross-account event handling via EventBridge and SNS alerting
• Integrated PyTest integration tests covering Django infrastructure routes as CI/CD quality gate, blocking deployments on failure and reducing change failure rate
• Engineered Celery + RabbitMQ async task processing with multi-layer health-check endpoints, enabling proactive incident detection and faster recovery
• Deployed LLM-powered semantic search using Amazon Bedrock and S3-backed vector embeddings against production banking data, establishing a reusable multi-model deployment pattern adopted firm-wide
• Drove SOX compliance through automated testing pipelines, secrets audits, and failover runbooks

Sole-developer build of HIPAA-aligned healthcare platform with production GenAI infrastructure. Django REST, React/TypeScript, GCP Cloud Build, AWS Bedrock, multimodal document processing.

• Architected multi-model AI system routing across Bedrock (Claude 3 family + Titan) with auto-fallback chains and token budgeting, reducing projected monthly inference costs from $300+ to $84-165
• Engineered 12+ production AI prompts: chain-of-thought reasoning, few-shot examples, prompt injection guardrails for healthcare document processing
• Built real-time AI chat with 11 diagnostic tools via LLM function calling and SSE streaming
• Built multimodal document OCR pipeline extracting structured JSON from 14 healthcare form types with custom extraction prompts and schema validation
• Implemented computer vision system using Rekognition Video for Q15 monitoring; NLP-based PHI protection via Comprehend Medical
• Designed AI compliance engine using chain-of-thought prompting to evaluate documents vs. PA BHRS regulations + CMS guidelines with 4-category weighted scoring
• Architected HIPAA-aligned platform (Django REST + React) with 19-step Cloud Build CI/CD, canary promotions, and automated rollback
• Established PyTest-based testing infrastructure (209 test files) with mypy + django-stubs static type checking, enforcing type safety across codebase
• Engineered JWT authentication with RBAC, token rotation, blacklisting, and 7-year audit retention for HIPAA compliance
• Developed semantic search with pgvector embeddings (1024-dim) and audio dictation-to-data pipeline (medical speech-to-text to LLM extraction)