Sep 2023 – Oct 2024 · 1 yr 1 mo · Chennai, Tamil Nadu, India · Hybrid

• Led the End-to-End implementation of Generative AI solutions for a major Auditing and Accounting firm to help their practitioners compare, query and score the uploaded client ESG/CSR reports against the existing GRI, SASB, ISRS standards, ultimately boosting the effectiveness and efficiency of audit processes.
• Built “AI- Assisted Scoring system” and “GenAI Driven Sunburst Chart View” that delivered AI-generated responses, citations, and highlighted text within PDFs. Designed with offline RAG design for batch processing OpenAI responses and store them in DB, effectively reducing latency.
• Developed an “GenAI enabled - Insights Dashboard along with Chatbot QA” to extract pre-defined key metrics for client doc and initiate conversations with system.
• Oversaw a team of 13, leveraging technologies such as Azure Cognitive Services, OpenAI Embedding, OpenAI GPT-4, and Python

May 2021 – Aug 2023 · 2 yrs 3 mos · Chennai, Tamil Nadu, India

• As Lead Data Scientist at Wipro, I was part AI CoE Practice team, We successfully launched a global-level DataRobot MRM GRC solution for BFSI clients, I was managing a team of 6 Data Scientists and 5 Data Engineers while working closely with the Product and Delivery team to ensure seamless integration and delivery.
• We architected and developed critical use cases for a leading FMGC manufacturing client. We implemented demand forecasting that accounted for external events, such as news, using DataRobot’s multivariate time-series forecasting combined with text analysis. This approach enhanced pricing decisions and profit budgeting. Additionally, we optimized inventory for low-velocity products by employing DataRobot’s optimizer to balance product quantities in the warehouse, which minimized stockpiling and reduced working capital.
• In another project, we focused on refreshing a credit risk models within timeframe for a financial entity to avoid penalty by regulatory bodies, ensuring compliance with regulatory standards on a unified platform. We rebuilt and onboarded the existing models on DataRobot’s platform (VPC), incorporating Model GRC capabilities to adhere to regulations like GDPR, SR 11-7, and IFRS-9/CECL. This implementation included auto-compliance report, model versioning, bias detection, audit logs, model explainability, approval workflows, and tracking data and model lineage. It also addressed model decay through checks for data drift, concept drift, stress testing, and back testing, helped avoid penalties and ensuring robust model governance.
• For an oil and natural gas client, we developed an end-to-end NLP solution designed to predict and mitigate risks based on text observations noted by operators. By employing NLP Tagger and spaCy for named entity recognition, the solution enabled early identification of potential issues, which helped prevent unexpected asset failures and reduced overall costs.

May 2020 – May 2021 · 1 yr · Chennai, Tamil Nadu, India

• Developed a End to End 360-degree predictive analytics aimed to provide in the form of product as a service (PaaS), tailored for end-to-end insights across customer, marketing, and supply chain domains. This customized solution involved creating domain-specific analytical use cases and integrating them into a single platform. Key components included building OLS models to analyze sales versus channel spends, implementing a Marketing Mix Model, and designing proprietary architecture and algorithms using linear regression coefficients, ADBUDG, PulP, and demand forecasting techniques(ARIMA, SARIMAX, FBProphet). This holistic approach enabled rapid transformation of data into actionable insights, offering a streamlined solution that delivered quick value across multiple business areas.

Jul 2019 – May 2020 · 10 mos · Chennai, Tamil Nadu, India

• Rearchitected an existing on-premise Article Recommendation System by migrating it to a cloud-based Python stack, utilizing a hybrid filtering approach combining collaborative and content-based methods. This included implementing Tf-IDF, Word2Vec, matrix factorization, and cosine similarity techniques, leading to a 12% increase in customer conversion while reducing costs and maintenance efforts. Additionally, conducted an in-depth Article Performance Breakdown by segmenting user behavior patterns through RFV and Cohort Analysis, which provided valuable insights that enabled authors to create high-performing content, ultimately increasing customer engagement by 20%.

Aug 2015 – Jul 2019 · 3 yrs 11 mos

• Spearheaded multiple high-impact marketing analytics projects that delivered significant business outcomes. Developed and implemented a user segmentation strategy using the K-Prototype algorithm with PCA, which enabled personalized marketing efforts and boosted cross-sell and up-sell opportunities. Enhanced ad placement effectiveness by building a predictive classification model, resulting in a 40% increase in click-through rates and a 30% reduction in cost-per-click within a month. Additionally, designed and executed a Marketing Mix Model using OLS regression to measure the impact of marketing efforts, optimize budget allocation, and maximize ROI, leading to more effective marketing spend.

Oct 2013 – Jul 2015 · 1 yr 9 mos

• Led the creation of a centralized data pipeline on Azure Cloud for digital marketing data storage, aimed at enhancing reporting and visualization capabilities. The solution involved aggregating data from over 20 sources using Azure Functions, scripts, and Azure Data Factory (ADF) to efficiently move data into Azure Data Warehouse (ADW). This implementation provided multiple personas across different teams and lines of business with seamless access to integrated data, setting the foundation for future predictive analytics and data-driven decision-making.

Mar 2011 – Sep 2013 · 2 yrs 6 mos · Chennai Area, India

• Led the implementation of data acquisition processes using Talend and SSIS to transform data into structured sources for Redshift and SQL SERVER. Developed BI Dashboards using SSRS, SSRS, and Crystal Reports, leading to informed strategic decisions and 50% query performance improvement.

Jun 2009 – Mar 2011 · 1 yr 9 mos · Tirunelveli Area, India

• Developed a web application for online bus ticket booking at Tinkle Technologies, utilizing JQuery, CSS, HTML, PHP, and MySQL. Achieved a revenue pipeline with a profit of in lacs in 3 months, on-boarded 1500+ customers. Reduced human error and increased customer satisfaction.

Jul 2008 – May 2009 · 10 mos · Tirunelveli Area, India

• In my role as a Programmer at Periyar Computer Research Academy, I spearheaded the development of the E-Educare application and dashboard using ASP.NET and SQL SERVER. By automating 95% of manual tasks, we significantly improved operational efficiency by 80% and reduced manpower costs by 60%, resulting in a remarkable ROI of 6 lacs.