At IBM, I designed Ship Your Own Model in watsonx.ai so customers could onboard custom LLMs. I also drove LLM onboarding workflows that reduced onboarding time from days to hours.
I architected Jenkins pipelines to build and auto-update container images and LLM models, cutting build time from seven days to one. I also designed release orchestration workflows that reduced manual effort by 90%.
I led the migration of inference workloads to vLLM on OpenShift and managed LLM inference clusters across AWS and IBM Cloud. I resolved production latency and compliance issues and established monitoring and health checks across microservices.
Earlier, at FICO, I automated cloud provisioning and deployments with Terraform, Cloudify, Python, Jenkins, and Ansible. At SOTI, I built and maintained Jenkins pipelines integrated with JFrog Artifactory.
