Jun 2023 – Aug 2024 · 1 yr 2 mos

• Defining, executing, and building Gen AI application for enterprise operational use cases to augment humans for root causing incidents and reducing time to recovery.
• Built a proprietary data ingestion and RAG+LLM based system for knowledge base querying, tool calls, and Agent execution infra.
• Leading, mentoring, hiring, and driving execution of the overall engineering team.

Jan 2023 – Aug 2024 · 1 yr 7 mos

Nov 2021 – Nov 2022 · 1 yr · San Francisco Bay Area

• Leading Tech Leads and Staff/Senior engineers to design and build control planes for online databases, distributed consensus, and efficiency initiatives to shift the cost curve down via use of block store and data tiering across multiple AWS regions.

Dec 2019 – Nov 2021 · 1 yr 11 mos · San Francisco Bay Area

• Defining, designing, and building infrastructure for client telemetry data associated with Netflix product and playback experience on all user devices.
• Tech-leading effort to redefine data ingestion, processing, and serving, across product and analytics use cases to set us up for the next major scale up and new product explorations.
• Building asynchronous data processing architecture over Kafka, Mantis, and Microservices to power product analytics, real time observability, ML data ingestion, and serving real time materialized views.

Nov 2017 – Dec 2019 · 2 yrs 1 mo · San Francisco Bay Area

• Founding tech lead of RAS (https://research.facebook.com/file/4561236690664587/RAS-Continuously-Optimized-Region-Wide-Datacenter-Resource-Allocation.pdf) - a new system to dynamically guarantee capacity SLAs for all shared and private compute and data clusters across worldwide Data Centers for stateful and stateless services.
• Designed declarative heterogeneous capacity models for guaranteed and elastic needs. Built a control plane and services to iteratively integrate into the existing set of systems.

Sep 2013 – Oct 2017 · 4 yrs 1 mo · San Francisco Bay Area

• Defined, architected, and built a multi-region cloud native cluster management system for containerized applications, projects Mantis and Titus, now open source. Built control plane, API, and scheduler services for service, batch, and stream processing workloads.
• Author of open source extensible scheduling library Fenzo (https://github.com/Netflix/Fenzo).
• Founding engineer for Mantis and Titus.

Feb 2010 – Sep 2013 · 3 yrs 7 mos

• Architect and hands on technical leadership for large scale distributed resource, data, and system management software development for application in the grid/cloud/farm computing and Data Center optimizations.

Sep 2008 – Feb 2010 · 1 yr 5 mos

• Chief architect, delivering multiple generations of DReAM (Distributed Resource Allocation Manager) software and evolving the Microprocessor (cloud-like) Ranch with thousands of servers in multiple locations.
• Developed state of the art resource allocation and scheduling techniques for large scale compute clusters. Core features include advanced reservation, back-filling, job dependencies, deadlines, self tuning resource profiler, rapid scheduling, and scheduler tracing. These helped deliver consistent and sustained results of record utilization levels of the compute cluster while maintaining disparate SLAs among dozens of independent organizations.
• Hands-on technical leadership for resource and data management software technologies with 20+ member core team in multiple locations.

Oct 2004 – Sep 2008 · 3 yrs 11 mos

• Architect, technical lead for DReAM.
• Developed distributed system infrastructure for compute clusters. Core features included inherently distributed architecture for scale, fault tolerance, automated scalable deployments and operational management.
• Chaired meetings with users and stake holders including periodic, contentious cluster share negotiations. Created data aggregation, reporting, and tooling to help ease negotiations.

May 2002 – Oct 2004 · 2 yrs 5 mos

Oct 1997 – May 2002 · 4 yrs 7 mos

Jan 1996 – Oct 1997 · 1 yr 9 mos

• Design and develop data configuration management and version control system.