Senior Full Stack Software Engineer (FastAPI + SvelteKit)
October 2023 - Present
- Working with FastAPI and SvelteKit to build a new product from scratch.
Software Engineer (NodeJS Backend)
July 2021 - October 2023
- Leading the backend implementation, deployment and guiding frontend teams on key product features serving 100k+ users (Tech Stack:- AWS, NodeJS, ExpressJS, MongoDB, Redis)
- Achieved the quickest implementation and rollout of prepaid cards across India in just 4 months, as per Transcorp’s records which accounts for 1.4Cr Rs of GTV
- Implemented spam and scam detection logic using Bureau API which declined or deactivated around 11000 spammy users.
- Implemented cashback and rewards feature which increased retention by 25% for M2 numbers
- Implemented Gold cashback feature using SafeGold APIs for which 10% of prepaid card users signed up on the
first day
Software Engineer (Java Backend)
September 2019 - July 2021
- Designing, developing, and maintaining Design Time Layer for TIBCO BusinessWorks.
- Analysed and upgraded BusinessWorks to run on OpenJDK 11 which reduced licensing expenses by up to 1 Million $ annually for our biggest customers.
- Developed visual diff and merge – made the product the preferred choice for two organizations over competing options.
- Built templating engine and extensible custom icon support – facilitated customers by accelerating complex processes.
- Developed a JIRA story point-calculating chrome extension – that ensures saving a quantum of time in each instance of using the JQL view, especially for high-impact users such as seniors and managers.
- Fixed supercritical deadlocks in the product using JVisualVM.
Intern (Java Spring)
June 2019 - August 2019
- Developed features and fixed bugs in the company’s insurance software using Spring + Angular stack.
- Developed an optimistic locking feature across all REST controllers – applied Aspect-oriented programming, without drastically modifying the original codebase.
System Software Intern (Python)
July 2018 - December 2018
- Found new methodologies to benchmark power measurement of GPU, optimising different workloads to give consistent numbers used Python, Pandas, Matplotlib, NumPy and Jupyter regularly.
- Reduced variation in power measurement to under 1% from the original 6-10%, beating the initially targeted 3%.