Backend Engineer on WhatsApp Business AI

Building GenAI (LLM based) experiences for consumers and businesses on WhatsApp.

Backend Engineer on WhatsApp Payments Infrastructure.

Focus on the design and development of distributed systems solutions to onboard new payments use cases as well as improve the performance and reliability of existing ones. Some of the key initiatives I worked on so far in this role are,

● Drove design and development of a brand new multi-tenant service to offer server-driven State Machine as a service within WhatsApp.

● Led live migration of WhatsApp payments service to a low-latency, highly available distributed key-value storage system. As part of this migration, uncovered and resolved a critical race condition in the Payments flow which improve reliability and made the user experience more deterministic.

● Designed a highly reliable, distributed deletion framework for the deletion of data stored across diverse storage systems.

● Improved latency metrics of the service by identifying & removing unnecessary region hops.

● Developed a visualization tool for spatial inspection of Convolutional Neural Nets (CNN) using PyTorch, React & Flask.

● Built a Jupyter Notebook Extension using Jupyter Comms & IPython APIs, enabling CNN inspection within Notebooks.

● Migrated Facebook’s open source model interpretability library Captum.ai frontend from JavaScript to TypeScript.

● Developed an interactive web-based tool for visualization of neurites in light-microscopy images using ReactJs and spring web services.

● Extensively worked on HTML Canvas API for rendering and interaction piece.

● Improved Venmo’s performance through back-end code optimization, resulting in ~15% increase in payments/second.

● Developed a real-time monitoring system for a Machine Learning pipeline using Flask, redis and Server-sent events.

● Built a Machine Learning model using TensorFlow to predict payment traffic on Venmo, based on past traffic trends.