Narasimhan Venkadeswaran

As an Individual Contributor (IC)

• Led the end-to-end migration of the self-hosted Teleport Unified Access Plane for Domino to a cloud-hosted solution, ensuring secure access for both cloud and on-premises deployments.
• Designed and implemented Terraform modules for Teleport cloud deployment using CircleCI.
• Created Helm charts from scratch for Kubernetes client interactions.
• Developed procedures for Break Glass, audit logging, and granular RBAC within Kubernetes.
• Integrated alert systems with PagerDuty for proactive incident management.
• Produced thorough documentation while collaborating with multiple teams and conducting enablement sessions for the entire engineering organization.

As a Domino MLOps Platform Engineer

• Championed the adoption of DevOps culture through several key initiatives:Defined the "definition of done" for service reliability in the Data Model Images building pipeline, Model Building and Publishing stack, and version control of build model artifacts.
  Monitored the end-to-end Domino MLOps pipeline at both infrastructure and application levels using tools like New Relic, Prometheus, and Grafana.
  Built a Terraform framework to capture Data Model and Machine Learning user journeys, tracking state transitions of models at each phase.
  

As an Embedded MLOps Engineer for Domino Enterprise Platform

• Led by example and trained development teams in incident management and postmortem processes.
• Fostered engineering team participation in owning incidents and conducting postmortems, increasing developer ownership and improving the reliability of the Domino MLOps product.
• Reduced XL/Large incidents to Large/Medium at the 95th percentile and decreased monitoring ingestion costs by 30% through enhanced instrumentation.

Additional Achievements

• Designed and implemented end-to-end event routing capabilities within the existing IaaS framework for monitoring and alerting, utilizing Terraform for all Domino customers, resulting in a 30% reduction in the time to identify and resolve incidents automatically.
• Developed Python modules and libraries to identify issues with critical services, enhancing observability and reducing mean time to recovery (MTTR) during customer incidents.
• Improved the visibility and reliability of a critical Data Model Build component in the MLOps pipeline from 0% to over 75%.

Principal Production Engineer

• Owned and operated Yahoo's media content platform, serving 100,000 requests per second (RPS).
• Achievements include scaling the Cassandra backend to 30TB across five data centers and redesigning the platform in collaboration with engineering to reduce end-to-end latency from 1.5 seconds to 500 milliseconds, achieving four nines (99.99%) availability.
• Received the Yahoo Spot and U Rock awards.

Principal Production Engineer

• Managed the unified API middle tier running on Kubernetes, serving 50,000 RPS for Yahoo sites including Finance, Sports, and the Homepage.
• Technical skill sets include ingress routing on Kubernetes, service mesh implementation using Istio, and canary analysis using Kayenta.
• Actively mentored junior engineers.

Senior Tech Lead

• Led the Media Analytics Real-Time Pipeline utilizing Druid, Apache Storm, Apache Kafka, and API stacks to provide real-time analytics for content editors.
• Responsibilities included maintaining and upgrading the real-time analytics platform, ensuring monitoring and observability.
• Championed the adoption of Imply Pivot, a powerful visualization tool for real-time analytics.

Tech Lead

• Managed a private cloud based on Chroot jail for Yahoo media, hosting hundreds of frontend (Node.js) and backend (Java) applications.