Vignesh Sankardas Email & Phone Number

Remote

• Developing a diagnostic S-arima model that uses minute-level telemetry from millions of on-premise sensors to detect reliability issues across a 300+ MWh battery storage fleet. Model includes emailing alerts to site managers, and a dashboard enabling deeper analysis.• Help develop and maintain an internal analytics dashboard that provides data summary, analysis, and visualizations to internal stakeholders. Dashboard has significantly improved the response time and effectiveness of engineering team in diagnosing and resolving problems.

Texas, United States

• Designed and implemented a dynamic Airflow DAG, seamlessly automating the ETL process for daily Real-Time Market (RTM) reports. Leveraged MS Graph API for precise API integration, facilitating data ingestion into a SQL Server database.• Automated the daily revenue reporting through a Power BI dashboard that consolidates revenue data from Tesla and internal ADA systems.

Boston, Massachusetts, United States

• Implemented DEI analysis on 32k Boston University student's data for mining patterns among students of different ethnicities and evaluated the difference between first-generation vs non-first-generation students• Defined and tested hypotheses, conducted multiple t-tests, and chi-square tests on 26 different questions, recorded and presented findings to the Director of Boston University Newbury Center• Built a custom BERTopic model to categorise 100k+ exit survey comments by applying UMAP dimensionality reduction, clustering using HDBSCAN and deriving topics from texts using the class-based TF-IDF method

Boston, Massachusetts, United States

Course - CS550: Computational Mathematics for data analytics

Boston, Massachusetts, United States

• Discovered trends and characteristics of CEOs through in-depth analysis of Bloomberg data on M&A from 1999-2019 across GICS sectors, acquisition value and period• Constructed a multivariate regression model to accurately predict the survival rate of CEOs post-M&A utilizing 9 parameters

Boston, Massachusetts, United States

• Studied characteristics of medical expenditure in diabetes-related hospitalizations among 30k+ patients, identifying population segments, illustrating diabetes prevalence and executing t-tests on 12 subpopulations• Developed a regression model with 10 attributes to accurately predict hospitalization cost determined using IDC10 codes, achievingan accuracy rate of 93%.

• Described and implemented multiple heuristics for detecting bot accounts, efficiently capturing their activities across posts, transactions, and profile data, resulting in a significant reduction of AWS costs by over 35%.• Conducted an in-depth analysis of 18 user engagement metrics, including bounce rate, average session duration, and new vs returning user data from Mixpanel, utilizing SQLite and Python to identify strategies for reducing churn rate

• Created a pick check SQL stored procedure, resulting in a significant reduction of inventory mismatches, phantom allocations and false picks, resulting in an average labour savings of 1 hour per person, and a revenue increase of approximately $69,000.• Designed and implemented a heat map visualization of a Nestle warehouse utilizing SKU data and historical locations, calculated replenishment rates for over 400 products and performed ABC analysis to classify inventory items• Utilized SQL techniques to extract and analyze large sets of order data, utilizing predictive modelling through multivariate regression methods and K-Means clustering to accurately categorize order types and turnaround times with a 91% success rate.

Pune Area, India

• Analyzed data from SQL Server 2012 databases, Highjump to design an XGBoost time series model for predicting sales for over 300 retail stores with a 92% accuracy.• Constructed and presented a comprehensive inventory management dashboard using Power BI, utilizing data from multiple SQL databases, calculating key metrics such as turnover and quantity flow rates to improve visibility and control over the inventory.

Master Of Science - Ms, Applied Data Analytics

Course work : CS544 - Probability with R ; CS555 - Data Analysis and visualization with R ; CS566 - Analysis of Algorithms; CS677 - Data.

Bachelor'S Degree, Electrical And Electronics Engineering