Jul 2024 – Jan 2026 · 1 yr 6 mos

• Commerce is one of the activities that is uniquely human. Prosperity of a society has a positive correlation with the amount of commerce in the society (measured usually by GDP).
• At Simpl, we are rearchitecting how commerce should be done on internet. Coins are from BC era and credit cards are from the mid 1900s. How should commerce look like in the age of internet, mobile & AI? That's the question we seek to answer - and reshape the world as per this answer.
• I lead Engineering, Data Science and Risk teams at Simpl.

Feb 2022 – Jun 2024 · 2 yrs 4 mos

Mar 2020 – Feb 2022 · 1 yr 11 mos · Bengaluru, Karnataka

• I was the best of times, it was the worst of times!
• On the one hand I had an inside view of how a software company of tens of thousands of engineers having dozens of products used by billions of users works.
• On the other hand, inability to meet teammates & travel to the US due to COVID, coupled with the complex processes, tooling and organization at Google meant that I could not perform to the best of my potential.

Jul 2019 – Mar 2020 · 8 mos · Bengaluru Area, India

• Exploring, reading, exercising.
• I was figuring out some startup ideas, but could not figure anything out.

Jul 2014 – Jul 2019 · 5 yrs · Bengaluru/Taipei

• Built a team of ~25 people, which built a marketing automation platform offering android, iOS and web pushes. QGraph offered comprehensive solution around push notification uses cases like trigger notifications, segmentation, price drop alerts and personalisation.
• We got 30+ clients for QGraph from across the globe including from India, US, Europe, SEA and Middle East, raised an angel round of funding, and were finally acquired by Taiwan based Appier.
• Post acquisition, the team moved to Taipei and continued working on the product. We rebranded the product as AIQUA (https://www.appier.com/aiqua-overview/), and continued to enhance the product and sell it in new geographies.
• While in Taipei, I learnt some Chinese language (中文), honed my chopstick skills and ran and cycled along the river Tamsui (淡水). As of mid 2020, Several of QGraph members continue to be employed at Appier in Taiwan; and one has even found a girlfriend there!

Nov 2011 – Jun 2014 · 2 yrs 7 mos · Bangalore

• I worked on various projects:
• (i) Building a scalable server (comparing PHP/jetty/netty for the same). Figuring out various bottlenecks and resolving the same. Comparing how virtual servers compare with native w.r.t. cost efficiency of requests served.
• (ii) Led the Real Time Bidding team for around an year. Work on integration with exchanges, propagation of information (user visits, impressions, clicks, sales) to bid servers for better bidding, monitoring and alert systems.
• (iii) Started the machine learning project for smart bidding. Developed a model and employed it across a cross section of campaigns. Delivered superior results compared to default approach.
• (iv) Built the infrastructure for writing map reduce jobs in pig.
• Mentored engineers, helped management decide (tech) priorities, helped in hiring.

Dec 2006 – Jun 2008 · 1 yr 6 mos · Palo Alto, CA

• I did the following projects:
• Efficient migration of memory pages from a NUMA node to another NUMA node when the VM execution migrates across the NUMA node. This effort led to the following patent: http://patents.justia.com/patent/8719545
• A method to decide which guest OS pages should be backed by large system pages. This work led to the following patent: http://patents.justia.com/patent/8769184
• Schedule VMs across NUMA nodes to ensure fairness not only in execution times, but also in terms of locality of memory
• A detailed study of the pros and cons of the AMD RVI technology.
• Besides above, I was involved in hiring new people from the team, and guiding new joinees in their projects.

Oct 2005 – Dec 2006 · 1 yr 2 mos · Palo Alto, CA

• Analyzed performance of various workloads on 64 bit virtualization platforms. In particular, found answers to questions like following:
• For various workloads, which virtualization technique (the software i.e. binary translation, or the hardware i.e. VT) is a better technique
• For various workloads, which characteristics contribute to the workload running better with one of the virtualization techniques
• How can we write our code better to reduce the virtualization overhead
• How can the application be modified to run faster on virtualized platform