Feb 2018 – Jan 2020 · 1 yr 11 mos

• Instart Logic is world's first Endpoint Aware App Delivery Network. At Instart, we not only improve the performance of your App/Website, we also specialize in providing security against multiple kinds of attacks, be it DDOS, Scrapping, Gift Card Scrapping etc.
• One of the USP of Instart is to help site owners/web publishers to unblock ad serving on their site, so that their revenue stream remains intact and they can provide free/inexpensive experience to the end-user.
• India center of Instart is a full functionality center having all the functions like Engineering, Operations, Technical Solutions, Customer Support and Marketing.
• Engineering wise, the India Center specializes in developing security solutions and Big Data Analytics to provide value added analytical solutions to our customers.
• We are actively hiring in all these areas, especially in Big Data Analytics. If you are enthused by petabyte scale anomaly detection, pattern identification kind of problems and are good at BigData technologies e.g., Spark, Hadoop, Kafka etc., we have tons of interesting problems for you. Feel free to reach out to us.

Mar 2017 – Jan 2018 · 10 mos

• Instart is a App Delivery Network. In this role, I led the effort to define the security problem for Instart and coming up with a security architecture, which can solve customer's security concerns in SAAS model.
• Typical concerns for App and Website owners are scrapping, password guess attacks, gift card guess attacks, inventory blocking , DOS etc.

Jul 2014 – Mar 2017 · 2 yrs 8 mos · Hyderabad Area, India

• Amazon Logistics, like any other logistics provider faces issues related to Addresses and associated attributes. Addresses were designed for human consumption and are not particularly usable by machines. Geocoding is one way to solve the locatability of the address, but that is not sufficient. Amazon not only requires multiple usable-granularity geocodes, it requires additional information to make its last-mile operations efficient, cost-effective and agile. The additional information can be - if the address is business/residential, operating hours of the address, is it on a divided road, is it part of a high rise building or a stand-alone building and so on.
• My main contribution to this team was:
• To define the problem
• Come up with the concept of an unique identifier for each address in the world
• Map these addresses in an acyclic graph to use relationships like neighboring address
• Come up with 5 pronged strategy to solve the data issue
• Architecture of the system to acquire, mine and store the data for consumption
• Architecture of the system to vend this data for real-time use.
• Update: Since then, Google has launched a product called Google Plus codes (http://plus.codes), which solves same problem and they offer it as a service to other logistics providers.

Mar 2007 – Jul 2014 · 7 yrs 4 mos · Mountain View, Bangalore

• Infrastructure lead for GMail's Spam fighting team
• Developed next generation of reputation system.
• Created a generic infrastructure to track billions of timeseries in parallel, handling millions of update per second with a second level latency.
• Technical Lead for publicly available product Google Translator Toolkit - http://translate.google.com/toolkit (Being deprecated on 04-Dec-2019)
• https://blog.wikimedia.org/2009/06/09/google-translator-toolkit-supports-wikipedia/
• Conceptualizing new product ideas
• Design of products
• Implementation of products
• Ensuring deployment of products

Oct 2005 – Mar 2007 · 1 yr 5 mos

• At OpenClovis, I was responsible for delivery of High Availability related features (checkpointing, Health Check, Babysitting etc.) and Infrastructure features (Event Service, Notification Service, Naming Service, Distributed Lock Service, Distributed Leader Election Service)

Aug 2000 – Oct 2005 · 5 yrs 2 mos

Feb 1997 – Aug 2000 · 3 yrs 6 mos