Project Details

Transaction files arriving from multiple global regions often contain inconsistent schemas, missing fields and invalid records, making them unreliable for analytics. To address this, I built a serverless data pipeline on AWS that automatically ingests, validates and standardizes incoming datasets.
The pipeline is triggered when CSV files arrive at Amazon S3. A unified AWS Lambda function performs file-level validation, detects schema drift using configurable schema mapping and master schema files, and dynamically handles missing or newly introduced columns. It then applies row-level data quality rules including null checks, region standardization, transaction amount validation and currency-region consistency checks.
Valid records are written to a partitioned processed layer, while invalid rows are sent to a quarantine zone with error reasons. A DynamoDB audit table tracks processing metadata and data quality metrics, while Slack notifications provide real-time pipeline monitoring.

Project Details

Built an end-to-end cloud data pipeline during a data engineering virtual internship to help Atliq process finance and sales datasets for analytics. The goal was to automate ingestion, enforce data quality, and maintain traceability. I created a structured data lake in Amazon S3 and implemented an event-driven pipeline using AWS Lambda and AWS Glue (PySpark) to validate incoming files through schema checks, null validations, and row-level quality rules while isolating rejected records. To improve observability, I implemented metadata logging with DynamoDB and exported ingestion logs to S3 in Parquet format, enabling monitoring through AWS Glue Data Catalog and Athena. The validated Bronze-layer data was then transformed in Databricks using Apache Spark to standardize dates, handle missing values, normalize currency, and derive metrics by joining finance and sales datasets. The curated output was stored as Delta Lake tables in the Silver layer for reliable analytics.

Project Details

Fleetage Mobility require real-time visibility to reduce safety and operational risks caused by abnormal vehicle behavior. This project delivers a production-style streaming analytics pipeline that ingests vehicle telemetry data and performs low-latency anomaly detection.

Apache Kafka was used to capture high-frequency telemetry events generated by a simulated IoT producer, with Avro Schema Registry ensuring schema governance. Apache Flink processed the streaming data to identify overspeeding, low fuel, and overheating conditions and routed alerts into a unified topic. The alert stream was integrated with Azure Functions and Twilio to trigger automated driver notifications in near real time.

The solution demonstrates an end-to-end hot-path architecture optimized for scalability, fault tolerance, and real-time decision-making, while strengthening hands-on expertise in event-driven design, stream processing, and cloud-native data engineering workflows.

Project Details

Financial analysts at RBF (imaginary investment firm) rely on manual CSV-based workflows for daily market analysis, which delays insights and increases operational risk. To solve this, I designed and built a production-grade end-to-end batch data engineering platform that automates ingestion, validation, transformation, and analytics delivery for U.S. securities pricing data.

The pipeline uses Apache Airflow for orchestration, Amazon S3 for staging, and Snowflake as the centralized analytics warehouse with layered modeling across RAW, CORE, DIM, FACT, and Subject Area schemas. Incremental loads, data quality checks, and dependency management ensure reliable daily processing. Curated analytical views power Power BI dashboards for liquidity monitoring, sector analysis, ETF performance, and watchlist tracking.

The solution delivers consistent EOD data before market open, reduces manual effort, improves data reliability, and enables faster data-driven trading and risk decisions.

Project Details

ShopVista, a growing e-commerce business, faced challenges with data scattered across multiple files and systems, causing manual reconciliation, inconsistent reporting, and delayed insights. To solve this, I designed and implemented a centralized data platform on Microsoft Azure using Azure Data Lake Storage Gen2 and Azure Databricks.

Raw datasets such as orders, shipments, returns, customers, products, brands and categories were ingested into ADLS and processed using the Bronze–Silver–Gold medallion architecture with Delta Lake. Automated daily pipelines were built for dimension tables and fact orders, along with monthly pipelines for shipments and returns, applying schema enforcement, data quality checks and star-schema modeling.
The Gold layer delivered analytics-ready datasets optimized for Power BI, enabling interactive dashboards for sales performance, revenue trends, customer behavior and regional insights, while establishing a scalable single source of truth for analytics.

Project Details

CarePlus is an imaginary BPO company handling customer support for multiple clients, where support tickets and system logs were spread across different OLTP systems and dashboards. This fragmented setup limited visibility into customer issues, agent performance, and system health.

To address this, I designed and built a fully automated, serverless data engineering pipeline on AWS. Support ticket data was extracted from a MySQL database and system logs were ingested from cloud storage into Amazon S3 using raw and processed zones. AWS Glue and AWS Lambda were used to clean, transform, and convert data into analytics-optimized Parquet format. Amazon Athena enabled ad-hoc SQL analysis and curated data was loaded into Amazon Redshift Serverless. Power BI was connected to deliver unified dashboards for operational and business insights.

This solution eliminated data silos, automated reporting and enabled faster, consistent decision-making using a scalable AWS architecture.

Project Details

This project provides a Power BI Revenue Insights Dashboard for AtliQ Grands, a chain of luxury hotels in India struggling with declining revenue due to competition and static pricing strategies. Using five structured datasets (bookings, hotels, rooms, dates, and customer data), the dashboard uncovers trends in occupancy, ADR (Average Daily Rate), booking platforms, cancellations, and guest ratings.

Key findings reveal a flat pricing strategy, mismatched occupancy trends, and high dependency on third-party platforms. Recommendations include implementing dynamic pricing, enhancing guest experience to improve ratings, and promoting direct bookings via loyalty programs and incentives.

This project showcases how data-driven business intelligence can empower hospitality management to optimize revenue strategies, regain market share, and drive sustainable growth.

Live Interactive Dashboard

Project Details

This project is my submission for the Codebasics SQL Challenge, simulating a real-world data task for AtliQ Hardwares, a fictional global electronics company. As part of the analytics hiring process, I tackled 10 ad-hoc business questions using clean and efficient SQL queries on the company’s database. The analysis covered product growth, market performance, sales trends, discounts, and more. I also designed a Power BI dashboard and presentation to communicate key insights effectively to stakeholders. Through this project, I demonstrated my ability to translate raw data into strategic recommendations, apply advanced SQL techniques, and present findings in an executive-friendly format.

Project Details

This full-stack Expense Management System empowers users to manage their spending through a clean and interactive web interface. Built using Streamlit for the frontend and FastAPI for the backend, the system connects to a SQL database for secure data storage. Users can add, update, and visualize expenses across categories like rent, food, and entertainment, and analyze monthly trends. Key features include intuitive CRUD operations, category-based analytics, and time-based filtering. The project is modular and test-driven, with separate folders for backend, frontend, and automated tests. It also showcases REST API design, integration testing with Postman, and version control with Git. This system not only simplifies budget tracking but also encourages financial discipline by enabling data-driven spending habits.

Project Details

This project delivers a unified BI solution for AtliQ Hardware, a fast-growing electronics company struggling with Excel-based reporting. Using SQL and Power BI, I designed and developed end-to-end interactive dashboards across Finance, Sales, Marketing, Supply Chain, and Executive domains. The data model follows a snowflake schema with sources from SQL and Excel. I used DAX for complex calculations and DAX Studio for optimization. The solution replaces manual processes with dynamic, real-time dashboards and empowers the leadership to make faster, data-driven decisions. Key outcomes include improved forecast accuracy, cost control, and deeper insights across departments. The project also strengthened my skills in stakeholder communication, KPI alignment, and storytelling with data.

Live Interactive Dashboard

Project Details

This Excel-based analytics project for AtliQ Hardware, a fictional electronics company, demonstrates a dual-focus approach to data analysis — Sales and Finance.

On the sales side, I created interactive dashboards to track customer performance, market-wise comparisons, product-level performance (top/bottom 5), and divisional breakdowns. The goal was to uncover insights into customer behavior, sales trends, and business expansion opportunities.

On the finance side, I developed Profit & Loss (P&L) reports by fiscal year, market, and month, along with Gross Margin % reports by quarter. These helped evaluate financial health and supported strategic planning.

Technically, the project highlights ETL processes, Power Query transformations, relationship modeling in Power Pivot, and advanced DAX calculations. I focused on building user-friendly, insight-driven reports that support stakeholder decisions and long-term business goals.