• Led data analysis research, engineering design and release for a $500-800K workspace insurance revenue stream. Identified key requirements and rollout risks that influenced eventual software design.• Enabled a new company wide authorization system, was the first in the org to go live, with no downtime or rollback needed, providing early, continuous feedback to the authorization team. Impacted all paying customers at the company, and also allowed franchise owners to start using siloed services, helping realize a multi-million dollar opportunity.• Led test coverage initiatives for critical flows, that reduced wework.com site crashes, bringing the uptime to 3 nines for some pages. Led multiple redesigns and migrations of the internal hospitality site.• Mentored junior engineers, on code quality, arch flexibility and reliability but also on work communication and setting expectations and managing immediate stakeholders.

• Focused on different build engine components of the One Engineering System push at the company• Improved the speed and stability of the multi-level distributed caching infrastructure used by the build system• Worked on a newer, experimental version of the build system frontend that can onboard larger, complex repos that requires minimal developer effort while delivering massive gains from cloud builds and distributed caching.[github.com/Microsoft/BuildXL]

I develop and maintain an in-house distributed system to load balance http services. The product is built from the ground up, written in C++, deployed on unix, is asynchronous callback driven and multiplexes >10K connections on a single thread.Added a threaded logging component with knobs built in for performance tuning, wrote the project's build and deploy scripts with support for continuous integration. Ported complex business logic from a legacy load balancer, refactored code and added tests. I work on product rollout as well; aiming to retire the legacy product in slow, gradual manner.My role also involves fun tasks like investigating redundancy issues, failover, capacity planning and performance degradation(at consumer and hardware levels), and maintaining consistent state across distributed machines. I also develop and support SaaS infrastructure on unix and openvms. (more details on request)I am part of the oncall rotation on a team that aims for an availability of six nines.

I designed and wrote a metadata management service for a petabyte-scale wide area file system, by implementing a Directory Listing Service(DLS) in Java using berkeleydb as the caching metadata store. Added in support for FTP/SFTP/GridFTP and exposed RESTful services for web/android consumption using jersey API and forced scheduled jobs to run in background threads.I also ran performance benchmarks against the listing service and performed a comparative analysis against established, competing products.

1) Used C++, WIN32, COM, ATL to develop a document management system for windows desktops.2) Developed an IE toolbar to consume web services, implemented inter-process session management using WIN32 kernel objects.3) Created an online document management system using MS Sharepoint and managed deployment workflow.4) Handled HTML Email and Landing page development, created reusable xml templates