Hi, I'm Arpit
I take products from ZERO to ONE and beyond
Technical Architect
@Unacademy
Previously worked at
@Amazon
and
@Practo
You can find me @arpit_bhayani
Book I'm reading
I share highlights from the book as I read
Currently I'm reading The Fabric of the Cosmos and you can read all highlights here.
All books I've read
Nuggets
My thoughts on
I call them Nuggets and I think you’ll ❤️ it
Read'em hereLast blog I wrote
I write tech articles every month on various topics ranging from System Design, Databases, Benchmarks and Devops.
Read'em hereAn efficient way to stop an iterating loop
There are two ways through which we can stop an iterating loop, first by using break statement and second by making loop condition false. Let's see if one is better than the other.
Companies I've worked for
In 6 years of my work-ex I have worked for an early stage startup, a rocketship and a FAANG company. Here is the gist of work I did at each
Download my Resumé
UnacademyI ensure that we do things the right way and build systems that scales fast and well. Stuff that I focus the most is Search and Notification system.
AmazonI was part of the team that built e-commerce database of Amazon which meant I was always on my toe. I also wrote a data pipeline system which moved data at multi-million QPS.
PractoI did Platform Engineering at Practo and was part of DevOps which meant I managed infrastructure and made life better for developers by writing few services here and there.
Stuff I can help you with
Things I've built and scaled
Helps client mux multiple GET api requests into one and boost the user perceived performance.
Relevance on the fly, entity extraction, intent identification, related searches and query corrector engine.
Fans out sms, e-mail, in-app and push notifications to users in near-real time.
Rule engine for coupon creation, consumption and exhaustion.
Fault tolerant and reliable data pipeline system with ETL, multi-consumer and multi-sink support.
Automating mundane corrective actions to be taken when system needs it.
Ability to define workflow in terms of a graph where each step maps to a function to be executed.
Populates data in an vanilla staging instance by calling out APIs with faked and masked data.
The tool deployed python apps to over 150 servers in parallel with atomicity and auto-rollbacks as native features.
Rolling out feature to users incrementally, as defined by their geography, privileges, and other factors.