Ramijul – AWS, React, Typescript

IoT-based Smart Pads for senior housing, with features including python and Postgres backed data pipeline, and analytics on Zoho Analytics and custom React App (to address Zoho limitations)

• Built a data pipeline to process near-real-time IoT data
• Implemented data transformations on Zoho Analytics
• Built a client dashboard to identify Wet Beds in near-real-time using a React App integrated into Zoho
• Built a custom React App to visualize data (PoC for a potential client dashboard)

It is a trading market research platform that delivers near real-time market data and customized trend analysis through a custom-built multi-tenant CMS solution.

• Saved 90% in third-party API costs by optimizing API calls and pagination;
• Utilized scheduled tasks with AWS EventBridge and AWS ECS Fargate to sync over 2 million trading records multiple times a day using Node.js;
• Developed a serverless multi-tenant CMS solution using AWS Lambda, Next.js, GraphQL, Serverless Framework, and MongoDB;
• Built an analytics data pipeline using an event-driven architecture that scaled to process over 1 million events flawlessly in a 5-minute window, utilizing Node.js, NestJS, DynamoDB, RabbitMQ, and OpenSearch;
• Conducted an investigation that resulted in a 40% reduction in cloud expenditure;
• Participated in system design, code reviews, and documentation;
• Implemented a CI/CD pipeline using GitHub Actions, improving team efficiency, software integrity, and code quality;
• Enhanced service monitoring and observability with AWS CloudWatch.

A platform designed to generate optimal lineups for betting on external sites. It utilizes high-speed data pipelines built with Python and linear programming to provide near-real-time player statistics. These insights are processed by an in-house optimizer and ranking system, applying advanced mathematical techniques to enhance lineup selection.

• Designed, implemented, and released a SaaS MVP within one year, utilizing a distributed microservice architecture with an event-driven paradigm;
• Increased subscription retention by 35% by developing a Python-based high-speed data pipeline that generated real-time player statistics within a 1-minute interval, reducing subscription lapse from 45% to 10%;
• Led and mentored a globally dispersed team of three, fostering agile principles, which resulted in incremental improvements, service stability, and product-market fit;
• Implemented a robust CI/CD pipeline to ensure seamless integration and deployment using GitLab CI. Enhanced service monitoring and observability with AWS CloudWatch.

This project is the result of multiple iterations since 2017. During the global lockdown, a final pivot towards an Instacart-like business model—focused on delivering essentials within one hour—led to significant success. The e-commerce infrastructure integrates a nationwide network of stores and a fleet of freelance delivery personnel under a single platform, ensuring seamless daily operations. Custom-built software optimizes efficiency, streamlining logistics and order management.

• Transformed ShopHobe into DeliveryHobe;
• Drove the migration of a legacy system to a fully JavaScript-backed system in 3 months, all the while being hands-on with the implementation, utilizing microservice architecture deployed on AWS ECS Fargate;
• Reduced average order fulfillment time by 33% by identifying areas of optimization and automation using data-driven approaches.