Arpit Bhayani
I make code work at scale.
I'm a Software Engineer based out of Bangalore 🇮🇳, specializing in designing and building scalable backend systems; currently working at Unacademy as a Technical Architect.

Where I've Worked

February, 2018 - Present

Technical Architect

  • Lead the Search and Content Discovery Team
  • Designed and developed Search and Content Discovery platform
  • Introduced and developed Search Query Auto Completion
  • Bettered the result relevance and in turn improved CTR by 40.2%
  • Developed a service that identifies the intent from a search query
  • Designed algorithm for identifying Trending searches on platform
  • Designed Search Query Logging and Visualization system
  • Tech: Python, Golang, Elasticsearch, NLP, ML
December, 2016 - February, 2018

Software Development Engineer II

  • Developed a service that puts infrequently accessed data in cold tier
  • Created streamlined and seamless access for Hot and Cold Data
  • Bootstrapped a generic data pipeline platform for Amazon Retail
  • Developed a platform that enables auto-healing of services in despair
  • Tech: Java, DynamoDB, S3, EMR
September, 2015 - November, 2016

Sr. Software Engineer

  • Developed the tool that enables incremental release of products
  • Built the product that identifies anomalies in real time
  • Built the database as a service ensuring high scalability and availability
  • Built the tool used for dummy data population across products
  • Built the tool used for provisioning personal staging environments
  • No downtime migration of Staging Infrastructure to a cheaper Region
  • Tech: Puppet, Vagrant, Python, Flask, Boto3, Semantic UI
July, 2012 - July, 2013

Software Associate

  • Built Offline reviews feature in Saba's SPC Product
  • Implemented Language Check feature for Saba's Internal Tool
  • Tech: Java, Apache FOP

Some Things I've Built

Potterverse

Potterverse is a platform for Potterheads like me to discover something new, something interesting about the Wizarding world of Harry Potter. It will be powered by some sophisticated algorithms, Machine Learning and Natural Language Processing, to bring out details, facts and theories that are buried deep within.

Elasticsearch
Golang
Natural Language Processing
Information Retrieval

FlaskSR

FlaskSR provides an easy way to make flask pages load faster and better by streaming HTTP Responses. Every web application tries to minimize the "Time to First Paint". This can be done by streaming partial responses to client. FlaskSR does exactly that for Flask framework.

Python
Flask
HTTP Streaming