Rahul Salgare

Inventory risk score prediction

• Contributed to developing a predictive model that estimates the risk score of a store, predicting its likelihood of accepting an order if assigned
• I specifically worked on feature engineering, algorithm selection, and setting up MLOps, including building a data pipeline using Vertex AI, BigQuery, MLflow, and Kubeflow.
• The model is integrated into the Fynd order management system.

Jiomart

• I worked in the order management system team. Built different integrations for channels like JioMart, JioMart partners, Tira, etc.
• My responsibilities included leading the team, code review, and load testing.
• Large milestone: scaled OMS to handle 10 million orders per day.
• Designed and built a front-end-facing service for JioMart, which JioMart is using to limit the number of quantities customers can order per SKU/categories.

Embibe

• Worked on different small microservices like test generation, grading and assessment
• Built backend APIs using PyEve.
• Refactored the repository with better use of design patterns and SOLID principles.

Lathe machine monitoring

• Worked on developing a lathe machine monitoring system that utilized sensors to collect data on the performance of the machine
• Designed and implemented a sensor network, and worked with the maintenance team to develop predictive maintenance schedules based on the data analysis.
• Used the data to build the predictive analysis model using the KNN algorithm to classify the condition of machines into predefined classes.
• Implemented real-time alerts to notify operators and maintenance personnel of any issues that required immediate attention.