Jan 2023 – Sep 2024 · 1 yr 8 mos

• Led a comprehensive FinTech data project aimed at gathering, integrating, and analyzing job market data from various online sources. The project involved developing an automated data collection process using Python, Selenium, and other web scraping tools to extract information such as job titles, descriptions, salary data, posting dates, and locations from multiple job portals. I created and managed CI/CD pipelines in Git to automate these processes, ensuring efficient and consistent data extraction.
• Post data collection, I implemented entity resolution techniques to clean and standardize the dataset, followed by Natural Language Processing (NLP) to analyze job descriptions and identify key skills. This allowed the creation of a detailed title-to-keyword database, mapping job titles to relevant skills and median salaries. Additionally, geographical analysis was performed to assess regional salary trends and job market demands, with predictive modeling used to forecast job growth in specific industries. As the Senior Data Scientist, I led the project’s planning, execution, and team management, driving the success of the project through Agile methodologies.

Nov 2020 – Jun 2022 · 1 yr 7 mos · Bengaluru, Karnataka, India · Hybrid

• Led a data science initiative for an e-commerce platform aimed at improving customer experience by identifying and mitigating abusive behaviors from both buyers and sellers. The project focused on detecting fraudulent activities such as buyers returning items without the original product, sellers providing poor quality products, engaging in drop shipping, issuing fake tracking IDs, and causing long delivery times. By leveraging Natural Language Processing (NLP), the team performed sentiment analysis and keyword searches on customer comments to flag potential abusive behaviors.
• The project handled billions of transactions weekly, requiring advanced data engineering and analysis techniques. An entity resolution master dataset was developed to identify and flag new accounts linked to previously abusive sellers by matching parameters such as devices, payment methods, and addresses. Additionally, a pattern recognition system was implemented to detect sellers using fake tracking IDs. A dashboard was created in Tableau to maintain records of flagged sellers, allowing the legal team to collaborate on reviewing and confirming model results to avoid legal complications.
• The project improved customer experience by 2.3% and required the integration of various technologies, including Hadoop platforms, SQL, Python for statistical and NLP models, and Tableau for visualization and reporting.

Jan 2019 – Oct 2020 · 1 yr 9 mos · Bengaluru, Karnataka, India · On-site

• In this project, I assumed multiple roles, including training, technical oversight, and project management. My responsibilities involved developing and implementing a comprehensive training plan for new team members on technologies such as SQL, Python, and Tableau. I conducted training sessions, created a testing roadmap with assessment criteria, and evaluated the team members’ proficiency in these technologies. The training also covered problem-solving approaches, debugging and understanding existing and legacy code, and optimizing code to ensure team readiness.
• Additionally, I stepped in to replace a director-level stakeholder, where I led the creation of a master dataset in SQL, gathered from multiple data sources, and allocated resources for a Hadoop cluster. I also spearheaded the development of a web-based interactive tool with dropdowns and filters, enabling stakeholders to select and visualize their data, perform aggregations, and generate custom reports.
• My role extended to work allocation for the team, identifying and resolving roadblocks, managing the project timeline, and daily client consultations to ensure seamless project progression. The team successfully developed a reporting system to track key performance indicators (KPIs), eliminating the need for senior director-level oversight. Furthermore, we standardized processes across all teams and sub-teams, improving efficiency and consistency.

Sep 2016 – Dec 2018 · 2 yrs 3 mos · Bengaluru, Karnataka, India · On-site

• As a team lead, I spearheaded the creation of R and Python modules designed to serve as a comprehensive, plug-and-play analytics solution for the organization and its clients. These built to facilitate a wide range of data analysis tasks, allowing users to conduct in-depth Exploratory Data Analysis (EDA), regression analysis, binary and multi-class classification, forecasting, unsupervised learning, and text analytics without starting from scratch.
• One of the key modules I developed was an EDA tool available in both R Shiny and Python. This module connects directly to SQL databases or CSV files and provides a user-friendly interface for generating reports that include descriptive statistics (mean, median, mode), missing value analysis, univariate, bivariate, and multivariate plots, correlation analysis, and more. The tool also includes detailed explanations for each statistical method, enabling users to understand their data better and generate comprehensive reports efficiently.
• In addition to the EDA tool, I developed regression analysis modules in R and Python that offer step-by-step guidance through data wrangling, feature transformation, preprocessing, train-validation splits, and model building. The modules support a wide array of regression techniques, including OLS, stepwise, robust, LASSO, Elastic Net, Random Forest, XGBoost, MARS, and more. Users can also perform model comparisons using metrics like MAPE, MAE, RMSE, R-squared, and adjusted R-squared, with options for ensemble models and hyperparameter tuning.
• Other modules I led the development of include tools for binary and multi-class classification, forecasting models (e.g., Holt-Winters, Prophet, VAR), unsupervised learning methods (e.g., k-nearest neighbors, factor analysis, PCA), and a robust text analytics module.
• Furthermore, I worked on optimizing runtime by parallelizing processes and allocating resources across clusters in R and Python.

May 2014 – Aug 2016 · 2 yrs 3 mos · Bengaluru, Karnataka, India · On-site

• Gained hands-on experience with the Hadoop ecosystem, including foundational knowledge in HDFS, MapReduce, and basic data processing tasks.
• Assisted in the development of simple ETL pipelines under guidance, applying concepts learned in academic projects to real-world data.
• Participated in team meetings and contributed to discussions on data strategy and analytics solutions, learning the importance of data-driven decision-making.
• Engaged with senior engineers to understand the architecture of scalable data processing solutions, sparking a deep interest in big data technologies.