Piyush Yadav

Engineered a microservice leveraging LLMs to automate entity creation by auto-generating configurations. Reduced entity creation time by 50 percent by eliminating the need for manual configurations.

Created a service to provide filtered data by auto-generating SQL queries based on user inputs. Integrated with Databricks to fetch data within specified time ranges, boosting productivity by 20 percent for Product and Marketing Team.

Built a masking service using LLM-powered custom prompts to auto-detect and mask sensitive labels in images. Enhanced efficiency by automating language category detection for extracted labels, reducing manual workload and improving Customer Support Team productivity by 15 Percent.

Designed a service to periodically synchronize data between two microservices, addressing data loss caused by message queue errors. Eliminated the need for manual scripts and prevented RabbitMQ message pile-ups, improving system reliability and efficiency.

Designed Data Corrector Service whose Task is to eliminate all the Personal User data, hashed data, sensitive data and any invalid data using Pruning and various data filter custom algorithms.

Led the consolidation of three microservices into a single, modular architecture achieving a 30 percent reduction in infrastructure costs, a 20 percent boost in system performance, and a 60 percent decrease in codebase size, significantly improving maintainability and accelerating development velocity.

Created Adaptor Tool as a Microservice whose task is to automate the file upload process. The upload time has been reduced by 80 percent and file processing time has been reduced by 300 percent on average.

Orchestrated the migration of services to AWS ECS, optimizing deployment with zero downtime and automated production deployments.

Headed the Centralized Distribution Management System (CDMS), introducing features like Brands Sales Order Integration, warehouse management and enhancing code and query performance.

Designed a data pipeline queue using the Kafka Pub/Sub model, resulting in a 200 percent increase in data processing speed.

Successfully implemented Goods Receipt Note (GRN) functionality and integrated with Stock-One for managing inventory on daily basis.

Leveraged various AWS services including S3, EC2, ECR, Lambda, SQS, ALB, TG, and Task Definitions for a scalable infrastructure.

Collaborated with Salesforce to create Collection CRM microservice that increases productivity, reduces costs, and ensures the scalability of the collection team by 90 percent.

Created a Fraud Management System to identify users who repeatedly attempt to access our system with fake documents and block them from future use and containerized the whole service which helps in reducing AWS cost by 30 percent.

Increased 40 percent of productivity of team by deploying functionalities on the dashboard and automating manual tasks and build various dashboards for Product, MIS and Collection Team.

Created a website in Django for blogs and functionality includes Authentication and Authorization, update and delete the blog, Image add functionality, mail support. Tech Stacks : Docker, Kafka, ECS, ECR, S3, AWS Lambda, SAM, Redis, Node.js, HTML, CSS, Bootstrap.

Developed a power-saving methodology Tool by training a predictive model on various parameters, including humidity, day, month, and time, to optimize lighting costs. The model accurately predicts when the lighting component should be activated, resulting in efficient energy consumption and cost reduction. Tech Stacks : Python, Jupiter-lab, Matplotlib, Seaborn, Altair, Multiple Linear Regression.