Infrastructure orgTech Lead of Efficiency management frameworks team responsible for Capacity Based Efficiency Tooling (infra product)- In order to overcome scalability challenges related to Meta’s service capacity and expedite product releases, the primary aim of this tool is to optimize service capacity through enhanced efficiency. So our mission is to empower product and infrastructure teams to comprehend the factors influencing service capacity changes and guide them in taking action to achieve an optimal state of utilization.Tech stack : React, GraphQL, REST, Python, Hack, MySQL, Memcache, Messaging Queues, Hive, Presto, Profilers, Host Utilization metrics, Load testing, Capacity Estimation, Autoscaling on Meta’s Internal Cloud Infrastructure.

Demand and Efficiency management for Meta's Infrastructure

Hulu Ads Infrastructure org- Build the team with engineers and TPMs (Video, Ads, Infra) for Hulu 2019 hackathon to build Ad validation tools for our advertisers as the first step to our vision of Hulu self-service portal. We won the 2019 hackathon (#OneTeam), productionized this and received the patent in 2022.- I played a crucial role in Hulu Ads initiative, leading the Monolithic to Microservice decomposition. This involved implementing domain driven design principles, establishing a homogeneous set up, layered architecture for interservice communication and resolving latency bottlenecks.- Increased real-time Ad-serving performance by detecting memory leaks using JProfiler and benchmarking on different Java GC types (Parallel, CMS, G1) using Datadog/Grafana and AWS Cloudwatch metrics.- One of the founding member of ERG - HAPI (Hulu Asian Pacific Islander).

Storage Infrastructure orgCapacity Estimator Tool to estimate logical and physical capacity gain withnumber of drives added. UI driven Drive by Drive Expansion with Estimators :https://www.tintri.com/blog/2017/12/fast-and-efficient-drive-drive-expansion-all-flash-arrays

UI for monitoring Virtual Machines, Platform and Stats visualization on Tintri storage.

- training enhancements to the Neural Network based Self-organizing Map

- Enhanced VM Level Monitoring and Stats in VMStore by removing SOAP based calls and replacing it with REST based APIs using Java, JSON for several Virtual Machine Hosts from Vmware(vCenter), Red Hat Enterprise (RHEV-M), Hyper-V and OpenStack.

Call Center Tool Automation using Java, JSP and Servlets, Design Patterns and SQL where user data gets autodialed on the Screen after processing its quality through our scrubbing logic and Third Party Tools.