Umesh Yadav

Software Engineer

Bangalore, India

UY

About

Seasoned Software Engineer with 5+ years mastering Go, distributed systems, and high-scale applications. Adept at optimising for peak performance.

Work Experience

Lead Software Engineer

  • Design and setup API gateway for our financial api platform(zwitch).
  • Lead the team to design and build our savings accounts and current account services for payment platform to offer it to both our in house and outside vendor product.
  • Design and develop the common services used across the entire organisation to help reduce duplication. Worked on centralised KYC service, report generation service and in app inbox for invoicing.
  • Design and developed the entire lending product from scratch to generating 14million USD within 3months. This required leading a team of 7-8 engineers and collaborating with cross functional teams like product, design and bussiness.

Vedantu

2021 - 2022

Senior Software Engineer

  • Design and develop post-class doubt solving for the student. It handled 50k doubts per day and approx 200k messages between the student and teacher. This feature help increases the free to paid conversion by 15%.
  • I lead the new onboarding flow initiative across 4 teams. I am responsible for developing and aligning all the teams for changes required for the new onboarding flow. Also, I introduced the first Go service of Vedantu as part of this initiative. Service onboarded 50k daily users and handled 5k rpm. It improved the overall loading per page by 20x times compared to the previous onboarding service.
  • I lead the initiative of rearchitecting of core monolith service. Separated 3 services from the central monolith and moved the service<>service data sharing to an event-driven system.
  • Introduced team to CI/CD, unit-testing and clean code. Started daily release of services instead of 7-day release cycle.
  • Increased code coverage from 0 to 20% and set up ci rules to improve code quality.

Gojek

2018 - 2021

Product Engineer

  • Architectured and developed an in-house chat platform to felicitate driver<>customer and customer<>customer chat. The system served 1 million live users, 25 million messages per day, and 4 million new channels per day. This effort helped save gojek 6 million USD in vendor cost.
  • Designed and developed number masking for gojek users to allow them to make phone calls to each other without exposing their real number across 20+ different services(like food, ride, car etc) in 4 different countries. The system is serving more than 100k orders per minute and handling 4k concurrent calls per minute. We added automatic vendor switch to remove redundancy and make the switch as smooth as possible. This helped reduce the account take over cases by 90%.
  • Worked cross-team and collaborated with SMS, number masking providers to integrate it with gojek.
  • Helped design and developed push notification service which was used to send 3 million transactional, informational, promotional push notifications per hour.
  • Optimized OTP delivery for SMS to increase the user login conversion and reduce the overall SMS costs. SMS cost was reduced by 100k USD/month.
  • Lead the initiative to build an App Rating Widget to help improve the overall app rating of the service. This required building a real-time rule engine that decides when to show the widget and when not. This helped increase our overall app rating go up from 3.58 to 4.87.
  • Build an extension message platform for chat which allowed people to build a new type of extension messages independently to increase user engagement. Eg: image sharing, pay to friend/split, share restaurant, etc. This increased the order completion by 2%.
  • Build new communication channel on top chat to enable order related important message visibility. This helped educate our users and provide them with useful order related messages inside the order chat itself.
  • Build an automated KYC system for onboarding drivers on gojek. This reduces the overall onboarding process from 2 days to 1-2 minutes.
  • Designed and developed a centralized image service that was used to serve 80 million driver-related images like KYC documents, driver profiles, and other sensitive images.
  • Helped migrate ops platform service to Kubernetes. This required careful migration of services without impacting the users. This helped reduced the overall compute cost by 20%.

Education

University of Mumbai

2014 - 2018

Bachelor's Degree in Computer Science

Skills

Go
Postgresql
React/Next.js
gRPC
Kafka
Redis
Python
Docker
Kubernetes
AWS
Rabbitmq