Leaving Google
I recently left Google after five and a half years working as a software engineer.
It was a brilliant opportunity and in that time I learned far more than I might have imagined. A dream come true in many ways : materialized each step of the way. I met Sadhguru, I was able to connect with brilliant teachers, engineers, and peers, and I lived comfortably in California and across the world.
I spent my first two years there fulfilling a liberal arts anxiety of building real things that worked: designing complex configurable systems that could be adapted at a moments notice and still reliable enough to power “all the worlds business”. Outages were measured in millions of dollars and people were stressed out and in pain, designing themselves into midnight alarm calls as a form of job security. Wasn’t for me :) and I learned all that I could about robust and adaptable system design before asking for my next opportunity.
I spent the next three and a half years fulfilling another college dream : of building machine learning for Arabic language speakers. I was asked to improve Web Search Ranking (what comes first, second, and third in response to various queries) for Arabic users, and quickly became responsible for a range of language families that had particularities outside the Latin alphabet. This was awesome, for me, as I’d always loved languages and studied linguistics in school. I got to learn all the worlds language families, or at least the subset that had relatives working for Google :) with a particular focus on the Indic continent and all its beautiful alphabets.
One of the problems we faced with Arabic, beyond the tech debt and hurried design decisions that were easily, if painfully, remedied, was that google could only share what it could find. As more of the internet is becoming a walled garden, the internet is changing. The algorithm is designed, generally speaking, for the old web, and there is a new paradigm emerging, that the top didn’t seem interested to pursue beyond a moonshot design sprint that fizzled under internal resistance. I cared about that project — which would look like aligning a new search project with a more natural, visual-first experience, and I got to see firsthand how limited we can be, sitting in our self-reinforcing bubbles.
Everybody is chasing each other, and few people are taking the time to go outside and see what’s actually true for themselves. Especially in Silicon Valley.
With love, Razi