Aug 2018 – Mar 2019 · 7 mos

• Ola is India’s largest mobility platform and one of the world’s largest ride-hailing companies, serving 250+ cities across India, Australia, New Zealand, and the UK with ~3000 employees and generates ~$360 million revenue.
• Work experience:
• Setup data cluster for analytics community in Olacabs using Ambari as management tools which enable users to run analytical queries and ML jobs. Cluster supports Spark, Hadoop, Hive, Presto, etc.
• Worked on optimization/tuning and performance enhancement of Rstudio, ML clusters, and Big data pipelines.
• Setup monitoring and alert management on continuous/scheduling jobs on clusters.
• Achievements:
• Creating separate pipelines are clusters for analytical meeting reduce the load over ML clusters. Reduce ~50% overhead to jobs running on ML clusters in work time.
• Reduce overall running cost of ML/Analytical clusters by 40% using spot instances and also scaling/downscaling clusters on time as well as usage basis.
• Reduce the uptime of 100 nodes in ML cluster from ~1 hr to <5 mins using packer and initializing scripts written in python using Ambari rest APIs.
• Learnings:
• Treat your scaling plan like a business plan. Assemble a board of advisers to help you.
• Talk to the community. Don't hide and try to solve all problems yourself. Many brilliant people are willing to help if you ask.

Jun 2017 – Aug 2018 · 1 yr 2 mos

• Work experience:
• Responsible for prototyping, building multiple scalable in-house applications, and maintaining, sharing, and onboarding different teams in the company to use them.
• Sethji: Prototype and build from scratch in-house tool to calculate cost on different PaaS like AWS, Azure for making cost optimization decisions and team budgeting using Flask/Python as a tech stack.
• Fusion: Build from scratch in-house cost exporter tool which gathers information from different Sethji export it to Prometheus used Grafana as visualization tool using Go.
• ThirdUmpire: Build from scratch in-house tool to restrict the direct access to teams on different machines. They can run the required command from ThirdUmpire without login into the machine. Build using Flask/Python and HTML/CSS.
• Achievements:
• Using Sethji and Fusion for cost optimization and budgeting tool, teams were able to cut ~23% of the total cost on different PaaS.
• Reduce ~92% error duties coming due to human errors when having direct access to the machines.
• Learnings:
• Build it yourself. I spent a lot of time trying other people's solutions that just almost seemed to work, but not quite. It's better to build some things yourself so you at least have some control and you can build in the features you need.
• Denormalize a lot, Avoid complex joins, Avoid scanning large sets of data: Always store information after checking these three things.

May 2017 – Jun 2017 · 1 mo

• It's a leading provider of digital transformation and operational intelligence solutions for financial enterprises. Work as a service company for customers like Adcuratio, Standard Chartered, Aditya Birla, etc.
• Work experience:
• Build AMS platform (Advertisement Management System) using Flask/Python, ReactJS, MySQL, and HTML/CSS as a tech stack.
• Worked on generating reports for their partners to switch advertisements based on bidding.
• Learnings:
• Long-running processes should be abstracted to daemons, separate advertisement slot bidding confirmation/report generation software as daemon/other processes.
• Optimize the database. Use explain command in MySQL to check how your queries are running. Indexes may not be being as you expect.

Feb 2017 – May 2017 · 3 mos

• "LendIn" in Kuliza is the flagship lending product recognized by Deloitte and Nasscom as a new-age lending system for banks and institutes.
• Work experience:
• Build authorization and authentication system for LendIn which has multiple third-party integrations like Google, SAML, Oauth, etc using Django/Python, MySQL as a tech stack.
• Worked on signup/signout pages in LendIn using ReactJS as a tech stack.
• Learnings:
• Test everything. You want to know when you deploy an application that it will render correctly.
• Log everything. When the system is not performing as expected it will help you pinpoint the bug/problem.

Jun 2016 – Aug 2016 · 2 mos · New Delhi Area, India

• It's a fast-growing software company specialize in A/B testing, Web Optimization, Behavioral Targeting, etc. With customers like Microsoft, Lenovo, Walt Disney, Target, Ubisoft, etc.
• Work experience:
• Create a module for customers to invite other users on their testing campaigns with specified permissions using Larvel/PHP as a tech stack.
• Build a browser extension to track/create the goals for customer's campaigns on Chrome and Firefox.
• Reduce the latency for the VMO plan layer Rest APIs using different and reliable caching strategies and database optimization.
• Learnings:
• Make your application easily partitionable from the start. Then you always have a way to scale your system. Look for opportunities to break the monolithic application into microservices.

Apr 2016 – Jun 2016 · 2 mos · Gurgaon, India

• It's the fastest, easiest, and most reliable on-demand medicine delivery service located in Delhi and serving the whole of Delhi/NCR.
• Work experience:
• Build IMS platform (Inventory Management System) for medicines on a wholesaler basis using Django/Python, Postgres DB, and HTML/CSS as tech stack.
• Major tasks to solve inconsistent tracking of stock and inaccurate data at any given moment handle by creating centralize tracking and assignments of medicine per order optimally for partners to avoid inventory loss and reduce human errors.
• Learnings:
• Most performance comes not from the coding languages, but from application design. There are some cases where coding languages can be choose for better performance but that's not the case most of the time.

May 2015 – Aug 2015 · 3 mos · Noida Area, India

• It's a service company that provides vehicle/delivery tracking end-to-end solutions in website development and mobile app development.
• Work experience:
• Build CRUD operations (Create, Read, Update, and Delete) for location service using MERN ( MongoDB, Express, Node, and ReactJS ) as tech stack.
• Major tasks for me to fetch real-time locations handled having two documents in DB ( history and current ) to store locations. Updates current location whenever there is an update in the history document.
• Learnings:
• Think of other creative ways to get the same result. It's not that optimized but works like a charm in that situation.