Managing the Traffic Targeting team within the Network Infrastructure organization. TL;DR - Team is responsible for a suite of tier-0 services which help "assign" Meta app users (FB, IG, WhatsApp, Threads, Horizon, Cloud Gaming) to their best possible Edge Location in ~real-time operating under constraints imposed by user-generated demand, available capacity, network topology, ISP's Traffic steering preferences, Edge location's health and availability and user-experience telemetry. Think - Bin Packing, on steroids!

Traffic Infrastructure Organization. Led a team of software engineers to deliver the following: 1. Capturing and aggregating telemetry (latency, availability, bandwidth) signals from users of Meta apps towards Meta's Edge Points of Presense (POPs) and network appliances (MNAs) embedded within ISP networks. Producing aggregated telemetry data for consumption downstream by Targeting (Load Balancing) services. 2. Collecting server-side Transport level stats on retransmissions, packet loss, goodput, etc. from Meta's huge fleet of reverse proxies running on POPs and MNAs and exporting them to real-time streams for downstream consumption/aggregation. 3. Created a simulation environment to run parts of the Targeting (Global Load Balancing) pipeline - outputs of simulation used for Edge Capacity planning/forecasting and E2E integration testing. 4. Disaster Recovery Analytics service -produced user-readable reports of egress drops, user experience drops during planned and unplanned outages in Meta's Edge infrastructure.

Technical leader for the Elastic Load Balancing organization, which includes 3 of the most widely used and successful AWS services (Application, Classic and Network Load Balancer) which together account for a significant percent of total AWS traffic, which host some of the most popular domains on the internet, which run mission critical workloads on millions of EC2 instances for 100s of thousands of customers across dozens of industry verticals in 25+ AWS regions and 100+ availability zones. All 3 ELB products are natively implemented in EC2, they integrate seamlessly with a variety of AWS services - CloudWatch, S3, WAF, Lambda, Auto Scaling, ECS and many more. We operate one of the largest cloud fleets in the world on behalf of our customers. Key projects: 1. Lambda-as-a-Target for ALB2. User Authentcation on ALB3. Advanced Request Routing on ALB4. SNI support and smart certificate selection for ALB5. Several platform level improvements for better visibility, operational performance and improved agility with the organization.

Focused on technical leadership within the ELB data-plane team. Was part of the initial team that designed, implemented and launched the new Application Load Balancer product in August 2016. Key projects: 1. AWS Application Load Balancer, a new Layer-7 load balancing platform which supports HTTP2, websockets. Offers better performance and availability when compared to the Classic Load Balancer2. Web Application Firewall for Application Load Balancer3. Content Based Routing on Application Load Balancer

Design and develop exciting features for the Classic Load Balancer (https://docs.aws.amazon.com/elasticloadbalancing/latest/classic/introduction.html). Key projects: 1. Access Logs support for Classic Load Balancer2. Perfect forward secrecy support for TLS connections on CLB3. Connection Draining support for CLB4. Several architectural improvements including seamless binary upgrades during deployments, improved operational visibility, improved availability through Route53 health checks

14-740: Fundamentals of Telecommunication Networks with Prof. Rajeev Gandhi, Senior Systems Engineer, Electrical and Computer Engineering Department, CMU, Pittsburgh.

Designed and developed a Web Service that measures and helps reduce security and abuse related incidents in Amazon's public EC2 Network using data collected form a wide variety of external sources.

Worked on a hardware accelerator called PISA (Programmable IP Services Accelerator), which is capable of performing state-full Deep Packet Inspection at very high speeds to classify TCP, UDP and IP traffic flows, apply rate limiting policies, access control policies, collect and export netflow stats, etc. This module was developed for the Catalyst 6500 switching platform, one of Cisco's flagship products in the enterprise switching market.Also, worked on a Quad-Supervisor redundancy system which could achieve greater Availability when compared to its predecessors. I worked on supporting the In-Service-Software-Upgrade module for this Quad-Sup configuration, along with other modules like active->standby switch-over, traffic splitting, etc. This was also developed for the Catalyst 6500 platform