Founder, Struct.AI (Jan 2023 - now)
I’m building Struct Chat — a thread-first, knowledge-rich chat platform, which blends real-time chats with threaded-conversations. Struct is a great replacement for Slack, Discord and Discourse. See announce blog post here.
This problem has bothered me since 2016. It is outside my comfort zone of a systems problem, but allows me to channel the product focus I’ve gained over the years, into solving a problem that affects millions of people every day.
0xFast (Feb 2022 - Jan 2023)
I built 0xfast.com, a 10x faster blockchain API platform, compared to Alchemy, Infura and various other players.
As my understanding of the crypto field grew, I lost faith in the field and decided to stop my work.
In my opinion, Crypto has fundamental issues. Bitcoin and Ethereum have failed to replace traditional currencies for any real-world use case. Lacking any real-world usage, the whole field has devolved into coin speculation, fueled by aggressive marketing.
Dgraph Labs (Oct 2015 - Jan 2022)
I was the (solo) founder and CEO of Dgraph Labs, Inc. from Jan 2016 to Jan 2022. I grew the company from 0 to 1M ARR, and from 1 (myself) to 50+ people distributed across US, India, and elsewhere.
Our main project was Dgraph, a Spanner of graph databases — geographically distributed, ACID transactions, with a GraphQL-like query language, powered by a graph engine. Dgraph was inspired by my work on Google’s knowledge graph.
Dgraph gained 19.5K GitHub stars becoming the most popular open source graph database on GitHub, surpassing existing graph DBs like Cayley, ArangoDB and Neo4j. Some of its most famous customers included ‘the top electric car manufacturer’, ‘the top shoe e-commerce platform’ and ‘the top health tech company’.
Dgraph Labs was the youngest company to be featured by Forrester in it’s graph database landscape as a contender, and the only native graph database to have been Jepsen-tested for transactional consistency and correctness.
Caching Dgraph was Ristretto, a Go alternative of Java’s Caffeine — a high performance LFU cache. Ristretto gained 4.8K GitHub stars and is most famously being used by Vitess and PingCap.
Integer compression, storage and communication was done via Sroar, a memory-wise cheaper and 6x faster implementation of Roaring Bitmaps — used to store sorted arrays of integers in Dgraph. IMO, it’s the most underrated project in all of my projects.
Looking at star power, the projects I authored and lead gathered over 37K GitHub stars, with each project exceeding in its category (two DBs and a cache).
Google (2006 - 2013)
Before Dgraph Labs, I was a software engineer at Google, where I co-led an effort to build Google’s graph indexing and serving system. Google’s Knowledge Graph System and OneBoxes (weather, movie showtimes, flights etc.) are utilizing this technology.
Before knowledge graph, I received an OC award for successful launch of Caffeine: Google’s incremental indexing system providing 50% fresher results, and the largest collection of web content ever offered by Google. For Caffeine, I built various solutions tackling size control, index quality and page freshness issues in a system holding >100PB of data.
I’ve lived in 5 countries
I grew up in Northern India and left when I was 18. I graduated with a Computer Engineering degree from Nanyang Technological University, Singapore in 2007 — my education fully sponsored by the Singapore government, on a grant afforded to only 50 students from India per year at the time.
Following my internship at Google Mountain View in 2006, I joined them full time. I worked for 2 years at Google Zurich, then another 4 years at Google HQ in Mountain View, California. I lived in Sydney, Australia between 2014-2018, which is where I started Dgraph Labs.
Since 2018, I’ve been living in San Francisco, California, which my family and I lovingly call home.