Hitesh Yadav

Software Engineer

Hyderabad, Telangana, India

About

Experience

  • Google ()
    • Staff Software Engineer
      Mar 2026 - Present · 5 mos

      Cloud ML Compute Services - Optimizing ML training / inference optimizations on TPU / GPU and Open-Source stacks.

    • Staff Software Engineer
      Mar 2024 - Mar 2026 · 2 yrs 1 mo

      Building solutions for compliance on financial instruments.

  • Software Woodworker at Kotak Mahindra Bank
    Jun 2023 - Mar 2024 · 10 mos

  • Amazon (8 yrs 2 mos)
    • SDE3
      Oct 2019 - Jun 2023 · 3 yrs 9 mos

      Worked on extension of ML based auditing solution and completed the UI, backend and data processing jobs for the solution. With the help of an extended teams contribution the solution is able to cover 70% of overall auditing space. As part of the auditing solution and discussions with customer figured out opportunity to create an internal feedback solution based on the financial data to improve planning. Currently, working on creating architecture and integration for the same.

    • SDE2
      Jan 2017 - Sep 2019 · 2 yrs 9 mos

      Moved to Auditing space as part of team change and worked on development of a system for auditing invoices from 3P carriers. Primarily owned payment authorization service and developed it's feature helped complete the launch with introduction of complete Dataware Housing solution. Created the warehouse using Redshift, Spark SQL and AWS Glue. Found opportunity of using ML for dispute resolution as part of the auditing team that allowed for faster dispute resolutions.

    • Software Development Engineer
      May 2015 - Dec 2016 · 1 yr 8 mos

      Worked on creation of a high scale invoicing platform solution for Amazon Transportation Organization. Primarily contributing to code, low level design and introduction of an ElasticSearch backed search solution for the platform. Additionally, worked on migration for migration for multiple service from SQL to DDB (key-value store). Found opportunities of cost savings by optimizing service call patterns and caching and reduced latency of a key service from 200ms to 1ms.