Work
  • Mar2026 - current
    Palo Alto Networks, Santa Clara, CA
    Staff Software Engineer
  • Aug2025 - Mar2026
    SellScale, San Francisco, CA
    Founding Software Engineer
    • Led the design and development of a metered billing platform using Python and React, enabling usage-based pricing and introducing three new pricing models using Stripe as customer volume and billing complexity increased
    • Re-architected a sequential campaign execution pipeline into a DAG-based parallel task executor, eliminating blocking dependencies and reducing end-to-end runtime from 9 minutes to 3 minutes under peak production load
    • Developed a RAG pipeline combining semantic retrieval and structured prompt injection, designing LLM evals to measure response accuracy and hallucination rates and improving multi-turn agents response quality by 35% compared to prior methods
    • Designed a multi-agent LLM workflow to automate real-time enrichment of large prospect datasets, reducing manual intervention by 50% while maintaining data consistency under concurrent updates
    • Led the development of a distributed async orchestration layer for bi-directional CRM sync using Celery and RabbitMQ, resolving race conditions caused by concurrent updates and reducing end-to-end sync latency by 35% under high-concurrency workloads
    • Reduced response-to-render latency by 60% by parallelizing dependent REST API calls and introducing client-side caching with React Query, improving load times across high-traffic dashboard views
  • Jan2025 - May2025
    Northeastern University
    Teaching Assistant
    • Working as a Teaching Assistant under Professor Marwan Subbouh for the Advanced Big Data Applications course.
  • Dec2024 - current
    Open Source Projects
    Contributor: Software Engineer
  • Aug2022 - Aug2023
    Sony, Bangalore
    Software Engineer
    • Scaled .NET app to handle 12 concurrent streams with optimized multithreading, achieving 10% faster processing times
    • Redesigned UI using WPF and MVVM architecture, supporting seamless display of up to 12 real-time video streams
    • Minimized frame drop rates by 20% by integrating RabbitMQ for real-time queue management from distinct IP streams
    • Achieved a 30% improvement in system reliability by migrating from MEncoder to FFmpeg to enable seamless playback and fault-tolerant processing for transcoding and broadcasting live streams via RTSP
    • Enhanced frame storage with asynchronous I/O and centralized database, achieving 10% reduction in data processing latency
    • Built a workforce productivity tool using microservice architecture with Java (Spring Boot), React, PostgreSQL, Docker deployed on AWS EC2 via Jenkins CI/CD, following TDD, improving team performance by 20%
    • Automated daily email updates by implementing an AWS Lambda function integrated with SES and RDS, saving approx. 10 hours per week and ensuring consistent, timely communication
    • Collaborated with the infra team to implement RBAC with Azure AD, enhancing security and reducing login errors from 50+ per week to less than 10, while streamlining user access management across the system
    • Developed alongside an Agile team of three and optimized RESTful APIs for key functionalities, achieving a 25% reduction in response times and 70% test coverage through comprehensive unit and integration testing

    link to project i worked on

  • Feb2022 - Jul2022
    Vrize, Bangalore
    Software Engineer Intern
    • Implemented advanced React UI components, like pagination, search and filtering boosting user engagement by 30%
    • Created automated test scripts with Selenium and Python, increasing test coverage by 20%, reducing manual testing efforts
    • Optimized front-end assets and utilized CDN services, improving website performance and reducing page load time by 20%
    • Developed Shopify Plus features, such as custom filters and enhanced checkout processes, boosting UX and conversions by 15%
    • Implemented PayPal and Stripe integrations, facilitating secure and seamless payment processing for 15,000 customers, reducing errors by 25%