Sanket Chavan Profile Image

Sanket Chavan

Senior Data Analyst

Data Analyst/Engineer with 5 years of experience in banking and IT, specializing in SQL, Python, Data Warehousing, and BI tools including Tableau and Power BI. Designed and implemented a fully automated compliance monitoring solution, tracking 130+ broker feeds in real time, eliminating 100% of manual processes, and reducing compliance officer data-processing time by 80%. Skilled at translating business requirements into scalable, data-driven solutions that enhance performance and inform strategic decision-making

Work Experience

Senior Data Analyst at Deutsche Bank

Nov 2020 - Present

Project: Employee Trading Compliance Applications (SQL, Python, Power BI, Snowflake)

  • Oversaw releases, capacity planning, and infrastructure updates for two enterprise-grade compliance applications as the Client Service Manager (CSM-1), ensuring high availability and operational continuity
  • Served as the key liaison for Compliance Officers and executive stakeholders, delivering BI dashboards in Power BI and producing ad-hoc data reports to support regulatory investigations
  • Designed and implemented SQL-based data validation scripts to monitor 131 feeds from 52 brokers, reducing incomplete feed risk by 85%. Integrated the process with ServiceNow APIs to enable real-time incident generation, accelerating resolution time from 5 days to 2 hours
  • Developed an automated reporting framework using shell scripting and Oracle stored procedures, eliminating 70–80% of repetitive compliance reports and reducing weekly L2 support requests from 150+ to ~40. Standardized output formats to improve downstream data handling
  • Created 13 KPI metrics in collaboration with compliance stakeholders to measure L2 team performance. Trained a globally distributed team of 16 analysts and scaled the dashboard across 18+ applications, reducing average incident resolution time from 5 days to under 24 hours
  • Engineered an Ansible-based access control automation pipeline that reduced privileged access incidents by 92% over two years. Logged all production activity to ensure full compliance and audit traceability
  • Designed and implemented a Snowflake-based data warehouse to centralize compliance datasets, currently in early phase. Built ETL pipelines and developed 20+ near real-time KPIs in Power BI to accelerate reporting and enhance visibility into employee compliance activities

Junior Data Scientist at People Tech Group

Sep 2019 - Nov 2020

Products : Elliptica Data Platform(Python, Kafka, SQL) & Candidate Screening Chatbot (Rasa, NLP, Python)

  • Developed fault-tolerant data connectors for both relational and non-relational databases to capture real-time data changes using Python and Kafka.
  • Streamed database changes to Tableau dashboards through Redshift Data Warehouse, enabling real-time data visualization.
  • Conducted stress testing on the data pipeline to ensure scalability across multiple use cases.
  • Collaborated with cross-functional remote teams for system design and sprint planning.
  • Led backend development for an AI-powered Candidate Screening Chatbot using Rasa and Python, integrated with PostgreSQL for data management.
  • Built a functional chatbot prototype within two months capable of understanding user intent, extracting entities, and delivering context-aware responses.

Data Science Research Assistant at University at Buffalo

Dec 2018 - Sep 2019

Research Project : Spatial-Temporal Analysis on GPS data (Python, Tableau, SSIS)

  • Spatial-Temporal Analysis on a Data collected using survey which is funded by National Institutes of Health (NIH).
  • Build ETL Pipeline to cleaned and ingest transformed data in MS-SQL using SSIS.
  • Constructed Interactive Maps for Visualizing trajectories of users and clustered those GPS points to locate the most visited places in R
  • Analyzed and visualized social influence on travel behavior of a single user and single household in R
  • Built the model to predict the next Activity of a user on a specific day & time using Markov Model and Conditional Probabilistic Model.
  • Designed a program to find Co-location between users using python and Visualize that Social Network in Tableau Dashboard

Data Analyst Intern at Niagara Falls Bridge Commission

Sep 2017 - Dec 2017

Research Project : Traffic analysis of Lewiston-Queenston Bridge (Python, ML, Excel)

  • Performed EDA on Bridge Crossing Data to discover Patterns and Trends in traffic
  • Conducted the Survey to identify the characteristics and travellers' behavior on change in toll cost
  • Achieved 13.11% traffic reduction during peak hours using the Multinomial Logistic Regression model

Projects

NYC Taxi Data Analytics with Azure Synapse (Python, SQL)

  • Developed and optimized SQL scripts and Spark notebooks to process three years of NYC Yellow Taxi data (2021-2023) in Azure Synapse Analytics
  • Configured dedicated SQL pools and Spark pools to manage and analyze over 100 million records efficiently.
  • Ingested, transformed, and loaded over 1TB of data using Serverless SQL Pool, Spark Pool, and automated pipelines in Synapse.
  • Enabled real-time analytics by integrating Synapse Link with Cosmos DB and visualized data through Power BI.
  • Analyzed payment behaviors, identifying that 60% of trips were paid by credit card, with Queens showing a unique pattern of higher cash transactions.
  • Developed Power BI dashboards revealing that Manhattan accounted for 40% of overall taxi demand, with demand peaking on Fridays and lowest on Sundays across all boroughs.

Movie Recommender System (Python, Pytorch)

  • Developed Item and User-Based Collaborative Models with Cosine, Pearson, and MSD similarity metrics, tuning and evaluating by measuring Hit Rates for recommended movies
  • Trained KNN and SVD++ models to predict user ratings for unseen movies in the test set
  • Achieved 75% accuracy in user preference prediction using a Restricted Boltzmann Machine
  • Enhanced recommendations with an Autoencoder model, maintaining an avg difference of 1-star rating between predicted and true ratings

Movie Revenue Prediction (Python)

  • Eliminated Anomalies using EDA, created 39 new features, imputed missing values using Prediction from BO XGBoost model
  • Built Linear Regression, Bayesian Optimized Random Forest, XGBoost, LightGBM models to predict the Revenue
  • Achieved an RMSLE score of 0.96

Volunteer Experience

Data Scientist I at VR Ulysses

Oct 2019 - Jan 2020

As the initial data lead at VR Ulysses, I focused on preparing and analyzing network data for their innovative virtual reality visualization platform, empowering security operators with enhanced insight and collaboration.

References

Dr.Qing He

Professor

I have had my pleasure of serving as Sanket's MS project advisor for Fall 2018, during which he analyzed large-scale smartphone GPS and extracted their social travel behavior. Above all, I was impressed with Sanket's ability to understand and solve the problem efficiently, his choice of visualization library to show his work was very good. He is also good at explaining his reports and finding clearly. I was also impressed with the way he designed the probabilistic model and tableau dashboard. Sanket would be a true asset for any positions requiring R, Python, and Tableau and comes with my heartfelt recommendation.

Prachi Sharma

Previous Coworker

Sanket's creative thinking, expertise, positive can-do attitude makes him an absolute pleasure to work with. He continually delivers results, goes above and beyond in providing exceptional work. He is very flexible, enthusiastic and hold the great interests towards technical domain. His strengths in staying across issues, pro-actively offering solutions and being adept at all aspects of communications make him a valuable contributor to team