Platform Team Planning & Data Issues Tracing System
Designed and built a full-stack data quality management platform to track and resolve data issues across the John Keells Group, enabling cross-business collaboration and faster resolution
Developed the backend using Django with RESTful APIs connected to Delta tables for real-time issue logging, status tracking, and cross-team visibility
Built a React frontend delivering an intuitive interface for issue reporting, workflow management, and progress monitoring
Engineered scalable data integration with Delta Lake to ensure reliable storage and efficient querying of structured issue data
Established CI/CD pipelines and deployed on Databricks Apps, ensuring system reliability, maintainability, and streamlined releases
0
4
Legal Document Assistant (Legal Sage)
Designed and built an AI-powered platform enabling users to securely upload and query legal documents through a fast, intuitive interface
Developed the backend with FastAPI and the frontend using React (Vite), implementing RESTful APIs, authentication, and multi-user document isolation
Integrated PostgreSQL (Dockerized) for secure document storage and metadata management
Implemented a Retrieval-Augmented Generation (RAG) pipeline using sentence-transformers for embeddings and Mistral 7B Instruct for context-aware responses
Engineered document preprocessing, chunking, and vector similarity search to deliver accurate, real-time answers from large legal texts
0
13
Weather Data ETL Pipeline
Designed and implemented an automated ETL pipeline to extract weather data from OpenWeatherMap API, transform it into structured format, and load it into AWS S3
Utilized Postman for API endpoint testing, authentication handling, and request/response validation before pipeline implementation
Deployed the entire pipeline on AWS EC2 with scheduled execution, demonstrating cloud infrastructure management and automation skills
0
23
Dynamic ETL DAG Generator
Developed a configuration-driven dynamic DAG generation framework in Apache Airflow (MWAA) to automate ETL workflows from Microsoft SQL Server to Snowflake, eliminating manual DAG creation overhead
Built modular ETL tasks for extraction, transformation, and loading with automatic dependency mapping and parameterized configurations, enabling rapid deployment of new data pipelines
Implemented incremental data loading strategies to optimize pipeline performance and reduce processing time