Feb 2019 – Oct 2022 · 3 yrs 8 mos

• There is software that is needed to make it easier to manage software and then there is software that directly solves a market facing problem. I am charged with solving the former at InMobi.
• The question that needs to be answered is what software is being created to enable, achieve and track system design concerns such as efficiency, reliability, scalability, maintainability, security, and other such usual suspects.
• Here is a sneak peek into the diversity of problems at hand:
• How does every developer on an average make 2-3 deployments to production in a day with a strong emphasis on safety and quality? Corollary: How does InMobi make > 20,000 deployments a year while maintaining the above properties?
• How does one design WebServices framework that is able to always run at a minimum of 80% CPU utilisation at all times, while still maintaining tail latencies within 40 milliseconds
• What does it take to embrace “Zero Trust” as the security design paradigm for InMobi?
• How does one manage logical data pipelines that have more than 5000 steps executing over a large number of data centers in way jobs always finish within tight deadlines

Mar 2012 – Jan 2019 · 6 yrs 10 mos

• The first part of my job is to solve for productivity of technology organization. As someone who has lived the life in an organization that had a total of 5 engineers, the interesting problem is around maintaining similar levels of agility in an environment that now has hundreds of technologists an continue to do so even when this number breaches the 1K mark. This involves periodic evaluation of how organic growth is impacting the conceptual integrity of various components and sub-systems, what the right set of abstractions and compartmentalizations are, and how things need to be shaped such that building on top of these basic blocks is easy for any smart engineer. The specifics might vary from something as mundane as looking API design to something a bit more involved such as extracting "plaformizable" pieces from existing applications.
• The second part is around harnessing the power of data. While the industry has an abundance of "big data" handling experts, the more interesting problem for businesses is to achieve convergence across what the data scientist, the hadoop expert and the business general managers' objectives. My specific role is around orienting everyone towards a common objective and then also be an integral part of the problem solving.
• Lastly, I continue to remain the core developer for a most of the internal server-side frameworks & libraries that are very demanding both in terms of scale, throughput & response times.

Sep 2008 – Mar 2012 · 3 yrs 6 mos

• Started off as the first software guy (besides the founders) with access to 2 servers and a team that was getting built. Led the overall system design to scale transaction volumes by a factor of 200 (that is 200 times, not 200%), made it run faster by a factor of 50 while using about 15 times more hardware.

Jun 2003 – Sep 2008 · 5 yrs 3 mos

• Started off as a engineer working on small yet complete products for finance. Briefly worked on infrastruture software. Then moved on to provide technical leadership for various new and old products like kids, cricket, weather etc. etc.