About Me
Hi, my name is Dhivya. I am a Microsoft-certified AI and ML Leader with over 10 years of full-time experience in Data Science, Machine Learning, AI, and Advanced Analytics. I have a proven track record in leading and growing high-performing data teams, backed by 3+ years of hands-on leadership experience. I have been successful in driving end-to-end data projects from scoping to building, and implementation, including A/B experimentation, monitoring, and maintenance. My key achievements involve spearheading the development of groundbreaking AI projects, such as a Creative Performance Prediction Tool using Deep Learning in Advertising, the worldâs first Interest Rate Deviation Engine using Machine Learning in Banking, and delivering the City of New Yorkâs first Block Sampling and Route Optimization Algorithms in four decades.
My specialties include: Product Data Science, Marketing Data Science, Fintech, Marketplaces, e-Commerce, Retail, Payments, Banking, Civictech and Start-ups
Skills
- Programming Languages: Python (Pandas, Scikit-Learn, Seaborn, Matplotlib, NumPy, SciPy, Jupyter, VS Code), R (R Studio, R Markdown, R Shiny)
- Machine Learning: Linear Regression, Ridge/ Lasso, Logistic Regression, Decision Tree, Bagging, Random Forest, AdaBoost, Gradient Boosting Machine, XGBoost, LightGBM, CatBoost, k-NN, Clustering Techniques, Bayesian Belief Network, Markov Chains, Recommendation Engines
- GenAI/ LLMs: LLMs, OpenAI - GPT, Anthropic - Claude, Google - Gemini, Search & Retrieval - RAG, Advanced RAG, RAG Eval, Embedding Models, Vector Databases, Prompt Engineering, LLMOps, Transformers, Hugging Face Transformers, Attention Mechanism, Perplexity.AI
- GenAI/ Image Models: Midjourney, Copilot, GANs, Diffusion Models
- Deep Learning: Artificial Neural Networks, Convolutional Neural Networks, TensorFlow, Keras, NLP, NLP Sentiment Scoring, Latent Dirichlet Allocation
- Data Engineering: SQL (Workbench, SQL Server), ETL, Data Pipelines, FTP, API Integrations, Python SDKs, Apache Airflow
- Azure: Azure AI Studio, Notebooks, Compute Instances, Blob Storage, Cognitive Services, MLOps
- GCP: Vertex AI, Vertex AI Workbench, Compute Engine, Tensorboard
- Databricks: Databricks Notebooks, DBFS
- AWS: AWS Sagemaker, EC2, Redshift, S3, Lambda, Glue
- MLOps: MLflow, Kubeflow
- Experimentation: A/B Testing, Multivariate Experimentation
- BI: Tableau, Plotly Dash, Power BI, Looker, ArcGIS, Carto
- Other: Salesforce Marketing Cloud, Adobe Analytics
Experience
The global financial technology platform that gives you the power to prosper. Products - TurboTax, Credit Karma, QuickBooks, Mailchimp
As a Senior AI Scientist, I build and evaluate RAG pipelines to deliver tax content at scale and velocity for Intuitâs Consumer Group using AI.
The Midwestern supercenter chain with 500 supercenters, grocery stores, neighborhood markets, and express locations across six states (MI, IL, IN, KY, OH, WI).
As the Lead Data Scientist at Meijer, one of the largest retailers in the US (contracted through Deel), I was part of the Meijer RMN (Retail Media Network) team, which collaborates with CPGs and brands to place advertisements on Meijerâs digital and in-store properties.
- Designed and built a Sales Attribution Model utilizing log and gamma distributions to measure campaign impact on in-store and online Sales.
- Researched and evaluated Google PAIR as a method to enable the Meijer RMN advertiser to deliver personalized and retargeted ads offsite to our audiences without relying on third-party cookies, preparing for a future without third-party cookie support.
Helping brands around the world make the most effective work, in the most creative way.
As the Staff Data Scientist, I am leading a team of Data Scientists and Data Engineers that deploys 260k+ CRM emails per month on the Salesforce Marketing Cloud and a portfolio of AI/ ML projects to bring together the power of data, technology, and creativity to build brands and unlock transformational and meaningful customer experiences for client accounts that include Edelman Financial Engines, LâOrĂ©al, AMC Theatres, SAS, UPS, Santander Bank, BMW, Circle and Veterans United among others at Performance Art, a specialist creative-data advertising agency and division of Interpublic Group (IPG).
- Pitched, won and built a Multi-Touch Attribution and Media Mix Model for Edelman Financial Engines to score the impact of paid and owned media channels, distribute conversion credit accurately, double attribution accuracy and simulate media mixes that optimize conversions/ ROAS on media investments.
- Pitched, won and built the SAS Fandom Score pilot, to algorithmically score music fansâ fandom using Spotify data and social mining.
- With DANCE, an AI product, I built a Creative Optimization Model to predict CRM creative performance before campaign deployment.
- With the Martin Agency, I worked on the UPS Impact Score Attribution Model to assess the attribution value for each marketing activity and find a performing marketing mix that maximizes ROAS and impact.
- Built a Sustainability Tension Maps product for Martin that uses NLP techniques to do Tension Discovery, Tension Scoring, Tension Ranking, Content Curation and Content Augmentation to bridge the gap between sustainable intention and action.
- Worked on the Santander Bank Wage Inequity Campaign to measure pay inequity.
- Designed a Leads Segmentation Model to improve leads quality and a Loan Recapture Prediction Model to predict loan recaptures after closing for Veterans United.
- Worked on the Circle Insights Engine to measure Brand Health including Brand Trust, Brand Loyalty, Brand Awareness, Brand Usage and Brand Efficiency for Circle.
- Worked on award-winning creative strategies to make SAS appeal to more Data Scientists - Superhero Popularity Prediction to create the ideal next superhero and Roblox New Game Design.
- Formulated strategies for BMW Munich for retraining the Next Best Vehicle Model and included recommendations that align with BMWâs 2030 sustainability goals like getting rid of the business rules that recommend owners vehicles of the same fuel type rather than BEVs and PHEVs.
- Worked on the Rolls-Royce pitch and designed the metrics to ensure RR remains a house of luxury using the RR House Proximity Score, RR House Affinity Score (similar to the CAS model for BMW), and the RR House Progress Score.
- Did vendor evaluations for procurement decisions for multiple vendors including Databricks, Shutterstock, InZata, Fivetran, CausaLens, Snowplow, Dash Hudson, Netbase Quid, Adverity, Hightouch, Meltwater, Meltwater APIs and Dow Jones.
- As a People Manager, I have a human-centered approach to managing people that includes forging strong relationships, unblocking blockages and avoiding micromanagement.
U. S. Digital Response
PRO BONO DATA SCIENTIST
Apr 2020 - PRESENT (> 4 years 10 months)
usdigitalresponse.org
We place experienced, pro-bono technologists to work with government and organizations responding to moments of crisis, to quickly deliver critical services and infrastructure that support the needs of the public. Weâre non-partisan, fast, and free.
- As an Advisor, I mentored and advised Project Leads on what it takes to deliver data and technology projects for Governments, on how to navigate bureaucracies, and on how to manage constraints and deliver tightly scoped projects.
- As a Pro- Bono Data Scientist, I built a first-of-its-kind immigrant services directory that centralizes services for 3 million immigrants living in Los Angeles County.
- As the Account Manager and Lead Scoper, I was involved with the creation of the technical scoping documents, establishment of strong relationships with the partners and placement of Product teams (Engineers, PMs, Designers and UX Researchers) for the City of Syracuseâ Innovation and Data Office, the City of St. Louisâ Department of Home Services and the City of St. Louisâ Elections Office.
The Only Platform you need to run Retail.
As the Data Science and Machine Learning Lead, I consulted at multiple data touchpoints within a San Jose, California-based retail startup- from Marketing to Product.
- Strategized and built a B2B marketing small business leads database with near-ideal accuracy using Google Places API and built a rule engine to weed out B2B leads that have limited propensity to convert.
- Built market insights to curate the marketing strategy for different customer segments in different geographies. Studied 27 verticals within the retail industry and identified new growth segments for the retail startup.
- Strategized merchant and validation DB rules to sanitize and standardize product data (including the product category, sub-category, product EAN/ UPC barcode, product name, description, image, etc) for seamless onboarding of new merchants and new products for existing merchants.
We use technology, innovation, AI, design, research, policy, connectivity and collaboration to make NYC future ready.
As an NYC[x] Innovation Fellow with the City of New York, I partnered with various city agencies to rapidly solve NYCâs problems with data.
- Partnered with the Mayorâs Office of Criminal Justice and developed the Hate Crime Index, a first-of-its-kind data product and algorithm built for NYC that scores the propensity of hate crime occurrences in the city. The index captures ~75% of all hate crime occurrences in the city by targeting ~50% of the police precincts.
- Partnered with the Mayorâs Office of Operations and built the NYC block sampling and route optimization algorithms for the NYC Street Cleanliness and Scorecard program. The algorithms facilitated and operationalized the collection of cleanliness data points for 17000+ additional blocks in the city annually, a 12x improvement for an effort designed to better the cleanliness-based quality of life for all New Yorkers.
- Built a Grant Funding Determination Engine using Machine Learning for the Department of Cultural Affairs, the largest municipal funder of culture in the US that determines $28 million in funds disbursed for ~2000 cultural organizations in NYC.
BankBazaar.com is the world's first neutral online marketplace for instant customized rate quotes on loans and credit cards. Shop for loans & cards just like you buy everything else now - online.
As a Senior Data Scientist, my job role had me build and manage projects at multiple touch-points within BankBazaar.Com- from the core product to operations and marketing.
- Built the worldâs first Interest Rate Deviation Engine using Machine Learning that predicts the best customers that can be made eligible for an interest rate cut for Personal Loans. The application rate of customers doubled with rate deviation and the approval rate improved by ~5%.
- Built Partner Preference Engines for Credit Cards and Personal Loans that predicts the customerâs âapplyingâ behavior and the bankâs âapprovalâ behavior. These Machine Learning models are used in the ranking and ordering of product offers and the offer position improved by ~19% for applied offers and by 22% for approved offers.
- Built Application Prioritization Engines using Machine Learning that enables for efficient customer life-cycle. This was done by predicting an applicationâs approval probability for Credit Cards and loan disbursal probability for Mortgages, Auto Loans and Personal Loans. The engines capture 80% of all approvals/ disbursals by processing just 50% of applications.
- Designed and tracked multiple targeted and personalized campaigns for Auto Loans using exploratory analyses. The campaigns drove an additional ~6% application volume to the existing Auto Loan portfolio.
- Enhanced customer data assets by mining unstructured data sources. The data were classified into different types of transactions using Natural Language Processing techniques.
- Built Credit Risk Engines using Machine Learning Algorithms for unsecured credit products that will assess the credit risk of serious delinquency associated with a customer for a particular credit product.
- Built a Product Recommender Engine that models on the customersâ historic online behavior data to predict the customers that had the highest propensity to apply for a credit card.
Grandatos.com is a simple initiative driven by Escomm Group, aimed at delivering Analytics insights, hind-sights and fore-sights for everyday use for everybody.
As a Chief Data Scientist, my job role had me pitch, win and manage Analytics projects for multiple clients.
- Built targeted segments with personalized content for multiple campaigns with unique threshold CTRs and the overall conversions improved by 20%.
- Built a Marketing Response Model that predicts the conversion of potential customers by employing direct email marketing.
- Developed Twitter Analytics models which help understand Twitter followers, analyze competitorsâ audience and follow #hashtag campaigns.
- Led the sales effort from Technical side by co-presenting sales pitches with customized data solutions and strategies.
IBM Global Business ServicesÂź is your business acceleration partner to co-create change and scale impact across your business. Work alongside our experts to reap benefits from connected processes, AI and automation, and a hybrid cloud architecture.
This was my first job after college. As a Software Engineer, I drove the effort in analyzing and reporting software quality for a Sales Automation tool built for AT&T.
- Appointed as the SPOC for all VoIP (Voice over Internet Protocol) related sales flow projects for both new and existing customers.
- Collaborated with SEs and Developers from multiple organizations, multiple cities and multiple time zones.
- Exceeded predicted targets for all projects and achieved recognition for having 0 red-flagged projects.
- Provided client training for Software Engineers on how to seamlessly transition from IBMers to IBMers partnering with AT&T.
Certifications
Azure Data Scientist Associates should have subject matter expertise applying Data Science and Machine Learning to implement and run machine learning workloads on Azure. Responsibilities for this role include planning and creating a suitable working environment for Data Science workloads on Azure, running data experiments and training predictive models. In addition, Associates manage, optimize, and deploy machine learning models into production. Associates with this certification have demonstrated knowledge and experience in data science and using Azure Machine Learning Studio and Azure Databricks.
Mentoring Experience
Sessional Mentor for the No Code AI and ML program at MIT.
Courses Mentored:
- AI Landscape
- Data Exploration - Structured Data
- Prediction Methods - Regression
- Decision Systems
- Data Exploration - Unstructured Data
- Recommendation Systems
- Data Exploration - Temporal Data
- Prediction Methods - Deep Learning and Neural Networks
- Computer Vision Methods
Sessional Mentor and Instructor for the PG Program in Artificial Intelligence & Machine Learning at the UT Austin McCombs School of Business.
Courses Mentored:
- Fundamentals of Artificial Intelligence and Machine Learning
- Supervised Learning: Regression (Linear Regression, Ridge/ Lasso)
- Supervised Learning: Classification (Logistic Regression, Decision Tree, Bagging Model, Random Forest, Gradient Boosting Machine, XGBoost, LightGBM)
- Ensembling Techniques: Bagging and Boosting Models
- Hyperparameter Tuning, ML Pipelines, Encoders
- Cross-Validation, Class Imbalance, Regularization
- Unsupervised Learning- KMeans and Hierarchical Clustering
- Artificial Neural Networks using Tensorflow, Keras
- Computer Vision and CNN
- Natural Language Processing
Education
Executive Program in Global Business Management (EPGBM)
IIM CALCUTTA
Indian Institute of Management Calcutta (IIM Calcutta or IIM-C) is a top business school in Asia located in Joka, Kolkata, India. It was the first Indian Institute of Management to be established, and has been recognized as an Institute of National Importance by the Government of India in 2017.
During my time at IIM Calcutta I secured âExcellentâ grade in 8 out of 12 modules. Courses:
- Financial Management
- Quantitative Analysis
- Operations Management
- Strategy Management
- International Business
Masters, Industrial Engineering
COLLEGE OF ENGINEERING GUINDY, ANNA UNIVERSITY
College of Engineering, Guindy (CEG) is a public engineering college in Chennai, India and is Asia's oldest technical institution, founded in 1794. It is also the oldest technical institution to be established outside Europe.
During my time at CEG I secured a CGPA of 8.91/ 10. Courses:
- Probability and Statistics
- Operations Research
- Simulation Modeling and Analysis
- Advanced Optimization Algorithms
- Design and Analysis of Experiments
Projects
Founded the data product, Maison ML to serve as a reference of data intensive COVID-19 reports to track the onset, progression and velocity of the 2019-2021 COVID-19 pandemic across the world. Maison ML has 260+ COVID-19 data reports hosted as of today. Tech Stack: The reports are powered by R Markdown running on R Studio Server installed in a AWS EC2 instance and knit into HTML web pages using the dynamic report generating engine, knitr. The visualizations are built using ggplot2 and the rich data frames are built using Formattable. The product is built using Jekyll and hosted on Github.
I make handwritten flashcards on Machine Learning and Artificial Intelligence concepts. These flashcards are intended to rapidly demystify complex Machine Learning algorithms and make them enjoyable to consume for non-data users. Flashcard Topics: - Scikit-Learn Functions - Linear Regression - Logistic Regression - Decision Trees - Bagging - Random Forest - Boosting - AdaBoost - GBM - XGBoost - Data Transformations - Regression Model Performance Metrics - Classification Model Performance Metrics - Regularization - Hyperparameter Tuning
Recommendations
âI have enjoyed the privilege of working with Dhivya for over 2 years. She is by far one of the most talented data scientists I have ever worked with. What sets Dhivya apart is not just her cutting edge approach to modeling and other machine learning applications, but her ability to explain what sheâs doing, why sheâs doing it and even how sheâs doing it, to the much less informed. As functions like data science are increasingly client facing, this skill is priceless. Dhivya is a powerhouse of knowledge and lives her passion for what she does. She is an incredible asset to any team or project. The icing on the data cake? Sheâs a lovely human to work with! ~ Janet Thompson, EVP General Manager IPGâ | âDhivya is one of the best data person Iâve worked with. Her ability to analyze complex data and transform them to actionable insights has helped me countless of times while working with her. Beyond that, sheâs also helped push the standard not just for our agency but also delivering future-forward thinking that pushes the industry forward. Dhivya also excels in high pressure situations with C-Suites both internally and externally. She would be valuable in any team sheâs in and cannot wait to see what she does next!â ~ Mitch Wong, Strategy Director IPG |
âI had the privilege of working with Dhivya as part of a pro bono project to design a map-based searchable directory of immigrant services for Los Angeles County. This tool is the first of its kind for an office of immigrant affairs. Iâm thrilled that in five months, Dhivya was able to help us conceptualize, develop, test and launch this critical tool for immigrant community engagement. Dhivya anticipated features and needs we didnât know we had. I look forward to the next opportunity to work with this data scientist and project manager extraordinaire!â ~ Michael Nobleza, FUSE Corps Executive Fellow, Los Angeles County Office of Immigrant Affairs | âIt was an absolute pleasure working with Dhivya. I supervised a project where she was a data scientist and developer. We were so impressed by Dhivya that we asked her to stay onto phase 2 of the project. Dhivya is thoughtful, adaptable, and very creative. She was able to seamlessly integrate her data and coding skills and create systems and even a dashboard to strengthen policy makers decision-making on various issues related to hate crimes, bias incidents and public safety. Itâs clear that Dhivya is a teamplayer, great communicator, and committed to public service and betterment of society. I hope I have the opportunity to work with her again!â ~ Hassan Naveed, Deputy Executive Director of The Office for the Prevention of Hate Crimes at NYC Mayorâs Office of Criminal Justice |
âIâve been working with Dhivya for more than a year now and never come across a peer who apply a structural approach for a problem and come up with an innovative and implementable solution as Dhivya. Strong fundamentals on statistics, love towards data & numbers and advanced skills in analytics tools always put her above rest of the peers. Her willingness to go extra mile, attitude on exploring new things and thirst for learning is exemplary and contribution towards improving the performance of BB product throâ analytics is fabulous. Itâs been a great to share a work place with such a superlative talent as Dhivya.â ~ Praveen Kumar Thandri Vijayaraghavan, Lead Data Scientist II - Decision Science and Analytics at BankBazaar | âI canât say enough about the problem solving skills that Dhivya brought to our projects. We had significant challenges and needed results while dealing with massive budget cuts due to the pandemic. Dhivya solved all our problems with open-source software and designed the custom tools to make our project shine. Her patience and thoughtfulness toward our project needs were amazing and matched very well with her overall knowledge and timely production.â ~ Gabrielle Stevenson, Outreach Manager at California State Treasurerâs Office |
âDhivya collaborated with our team from The NYC Department of Cultural Affairs and the Mayorâs Office to do a deep-dive analysis of agency funding data. Dhivya approached the project with a broad understanding of what we hoped to achieve and a deep curiosity about the minutae of our data. She created tools for us to identify trends in our data, introducing our team to new concepts and methods while showing us highy useful patterns and trends in our data. This project will be driving policy recommendations moving forward, and we could not have done it without Dhivyaâs creativity and attention to detail.â ~ Stacey McMath, Director, Programs, New York City Department of Cultural Affairs | âDhivya is an expert coder and she can quickly write efficient code for whatever be the analytics task at hand be it web scraping or machine learning. She has good command on current data science languages like R and is a dedicated worker/team player. All the very best Dhivya for all your future endeavours!â ~ Suniti Srivastava, Vice President - Machine Learning, Royal Bank of Scotland |
A Little More About Me
Alongside my interests in Data Science and Machine Learning some of my other interests and hobbies are:
- Acrylic painting and #GenAI art generating đ©âđš
- Home decorating đĄ
- Traveling and taking long road trips đ
- Watching Tennis đŸ