Building software systems for NVIDIA Cosmos: https://www.nvidia.com/en-us/ai/cosmos/

• Led development of a fully managed fine-tuning service.

• Developed benchmarking services to evaluate model performance.

• Developed services and tools to facilitate access to GPU clusters.

• Deployed and managed fleets of CI/CD runners.

• Built and deployed various AI applications across cloud service providers.

Developed features for LinkedIn's Feed, which serves more than 120K posts per second:

• Technical Lead for Reposts, which generate 1.5M posts per day and constitute 30% of content on Feed.

• Technical Lead for Collaborative Article redesign on Feed.

Developed project management and resource optimization systems (full stack):

• Developed a framework for asynchronous APIs (GraphQL/Resque).

• Developed various data models (MySQL), entity layer abstractions (Ruby/Rails), APIs

(GraphQL), and frontend features (React/Apollo).

• Introduced post-mortem practices that were adopted by the entire R&D department.

Developed control-plane networking systems for Google Cloud Platform - Compute Engine:

• Developed and launched Shared Virtual Private Cloud (VPC).

• Reduced backend programming for subnets by 98%.

• Developed project life cycle management systems.

• Developed data models, frontend service APIs, and backend workflows.

• Developed RPC clients to integrate with other services asynchronously.

Developed software systems to track Amazon's network topology:

• Developed a distributed application to discover and poll network devices.

• Developed an authoritative document datastore and secure API.

• Participated in rotating on-call duty to respond to service outages 24/7.

• Participated in organizational hiring and team on-boarding.