Gabriel Lewis

Spearheaded a multi-agent orchestration platform coordinating 12 production agents using LangChain, LangGraph, and MCP, reducing task completion time by 42%. Improved context routing with Redis (33% better recall), deployed scalable agent runners on Kubernetes with Docker and GitHub Actions (60% faster deployments), and integrated Databricks semantic layers to lift intent matching by 28%.

Migrated natural-language tooling to agent-driven automation for compliance workflows, deploying 8 agents that reduced manual review hours by 48% across a 200k record corpus. Built Databricks/Spark semantic retrieval pipelines to increase answer relevance by 36%, and engineered Kafka + Python microservices handling 1.2M events/day with 99.9% uptime.

Engineered secure enterprise API gateways and agent-facing endpoints with Java, Spring Boot, and REST, reducing integration time by 35%. Built OAuth2 token exchange flows for agent user delegation, designed PostgreSQL context models for multi-tenant deployments, and improved CI/CD with Jenkins/Docker (48% faster builds) while increasing release confidence via pytest/Postman contract validations (6

Deployed serverless agent connector prototypes on AWS Lambda and API Gateway, achieving 80% lower infra cost versus EC2. Built ingestion pipelines with DynamoDB and S3, implemented SQS/SNS asynchronous coordination (2x reliability under load), and improved monitoring with CloudWatch (44% fewer false positives) while automating environments via CloudFormation and early Terraform (provisioning from

Contributed to a Python agent connector prototype by implementing API clients and basic orchestration flows, reducing manual integration effort by 60% during demos. Benchmarked retrieval latency and implemented retry/backoff and structured logging to CloudWatch, improving transient failure rates by 45% and increasing core module test coverage to 78%.

Earned a Bachelor of Science in Computer Science from Texas Tech University from 2012 to 2016.