Hi! I’m Arjun,

An engineer and researcher currently working at Woven By Toyota. My key focus areas are personalization systems, data infrastructure, and machine learning.


Woven By Toyota

At Woven, I lead multiple projects in data infrastructure, for storage (operational stores, data warehouse) and processing (ETL pipelines, scheduling, collaborative workspaces). I also developed machine learning systems for things like adaptive assessments, and course recommendation systems as one of the founding engineers of MS1, a startup(ish) focused on talent management.


Previously, at BookMyShow, I worked on the development of real-time data pipelines for discovery and personalization systems, helping millions of users find entertainment experiences.


I did my undergrad in Computer Science from IIT Indore, which gave me a long-lasting curiosity in areas like:

  • algorithms (that learn from their data)
  • distributed systems (byzantine fault tolerant consensus algorithms, CRDTs)
  • language design (automatic performance optimization, program generation)


Growing up, I discovered websites like MIT OpenCourseWare and the Linux Documentation Project, which opened up a world of knowledge and opportunity. This experience shaped my belief in the power of open access to knowledge and technology to transform lives.

I am a big fan of companies and people that have taken a long-term approach and delivered results. Companies like Toyota, with their focus on quality and respect for people, Zerodha, with their no-nonsense approach to building a longterm focused business, and Charlie Munger’s interdisciplinary thinking and his ability to draw on diverse mental models to make better decisions.


I believe we are still in the early stages of algorithmic progress in deep learning systems. We have been able to make useful models by scaling up, we are yet to discover far more compute and data-efficient approaches. We are in the early phase of “Simplicity does not precede complexity, but follows it” cycle, and we will soon see a wave of simpler more cost-effective and performant models.

My current interests lie in the inference and training of large language models on smaller machines, mechanisitic-interpretability, and in-context learning.

I’m also fascinated by marketplaces and mechanism design, influenced by books like “Dataclysm” by Christian Rudder and “Who Gets What–And Why” by Alvin E. Roth I am interested in how we can leverage ML to create better marketplaces (reputation systems, matching algorithms, spam detection, pricing algorithms).


I love to read books, to better understand our world. Some books that I recommend are “7 Powers” by Hamilton Helmer and “Model Thinker” by Scott E. Page.

I also love biryani, mexican food, and BBQ. and enjoy skating and swimming.

Let’s Connect

If you’d like to discuss potential collaborations, job opportunities, or shared interests in machine learning, data infrastructure, or personalization systems, feel free to reach out. You can contact me at arjunsriva at gmail.