Sanjay Rakshit

About Sanjay Rakshit

Machine Learning Engineer, with around 7 years of experience in building scalable…

Experience

1. Senior Software Engineer - Machine Learning

   Freshworks

   Apr 2022 — Present · IN

   Learned PySpark, did multiple bug fixes for a large ETL pipeline. Identified schema corruption and added code to fix it. Actively fixed contact activity related feature engineering after collaborating with Data scientists- Added code to rate limit while pushing contact scores to downstream consumers. Added code to switch rpm dynamically between peak time and non-peak time- Adopted databricks and created jobs in an optimised manner to reduce the time taken for ETL and feature engineering- Single handedly restored and wrote orchestrator for account level refresh pipeline which was impacting contact scores and resulting in negative customer experience. This pipeline would gather all customer accounts that requires score refresh and proceeds with batch refresh. Also added pause and play feature while pushing scores. In case the job failed, we could restart it, and it would resume from checkpoint- Refactored the prediction code given by Data Scientists to fit industry standards. Realised that maintainability is as important as optimisation- Added a feature in real time system to avoid duplicate calls to a service, resulting in reduced cost. This helped reduction of external API calls by 50%- Created a chat bot service which would result in saving time of end users. This service, would help them lookup high intent contacts, generate introductory email, change the email as per their wish and then send it as well, all through the chat box. — This used Generative AI in the backend- Created a contact score decay pipeline, which would gradually decrease the contact scores as they would age without any positive activity. This helped end users to avoid contacts unlikely to buy- Added Jenkins CICD pipeline to our projects by collaborating with Devops, which would result to efficient development.