Harika Munnaluri

AI Initiatives for Salesforce Sales Organization: Leading a team of 7 data scientists responsible for crafting and executing the product strategy and vision for key AI projects, including Next best actions for sellers, AI Guided selling, Sales Pitch, Insights framework and experimentation platform

• Project SEED: Led the establishment and development of Project SEED, providing personalized next best actions to the Sales team.
• Strategic Product Management: Managed product roadmaps, aligning with company objectives and ensuring timely delivery of key features.
• Sales Pitch Generation: Developed a sales pitch tool leveraging Generative AI, including summarization and Retrieval-Augmented Generation (RAG) techniques, to assist sellers in pitching Salesforce products to customers.
• Competitor Analysis: Created a sentiment prediction model using Generative AI to analyze sentiment in call transcripts regarding competitors.
• Insights Framework: Utilized OpenAI to extract topics from deal notes, generating insights into deal win/loss reasons.
• Experimentation: Designed experimental platform to assess the impact of various product features on key performance indicators (KPIs).

Orchestrated multi-disciplinary team comprising data scientists, visualization experts, and data engineers to deliver cutting edge machine learning solutions for sales division at Salesforce. Some key projects include

1. Propensity to buy models: Developed predictive models leveraging customer data to assess likelihood of purchase. Incorporated advanced algorithms to analyze historical customer behavior, enabling sales team to prioritize leads effectively.

Achieved significant increase in conversion rates and revenue optimization.

2. Cross-Sell models: Engineered models to identify cross-selling opportunities within existing customer portfolios.

Resulted in enhanced customer engagement, driving incremental revenue through strategic cross- selling initiatives.

3. Predicting quota attainment: Designed and implemented machine learning models to forecast individual and team quota attainment.

4. Causal inference models: Developed sophisticated causal inference models to understand impact of various factors on sales outcomes.

DataRobot is an End to end Automated Machine Learning platform. Responsible for collaborating closely with the customers to achieve their objectives with AI

• Supporting 20+ customers spanning various industries such as Banking, Telecom, Retail and Insurance on their AI lifecycle by providing data science expertise and helping achieve quantifiable value
• Providing voice of customer to Product and Engineering teams
• Led implementation of various AI/ML use cases, encompassing Churn models, Risk models, Propensity models, Lead scoring models, Forecasting, Image recognition, and Fraud models, among others.

Responsible for managing the Initial Credit Limit models for International markets

• Leading team of data scientists for end-to-end development and implementation of credit limit assignment
  models using machine learning algorithms Revamped global Initial Line framework leveraging Ultimate Gradient Boosting models to predict spend and revolve balance of customer.
• New framework delivered incremental revenue of $30M in two years
• Designed Transfer Learning framework for Credit Limit assignment in smaller markets; thereby increased
  coverage of models across International markets
• Collaborated effectively with senior stakeholders in business, market teams, strategy and tech teams to ensure seamless implementation of models

Led end-to-end responsibility for requirement gathering, idea prototyping, daily client interactions, design, implementation, and delivery of advanced analytics products. Successfully developed and deployed machine learning models for property-related use cases.
Forecasting House Price Index: Constructed forecasting models utilizing a two-stage error-correction econometric model. Leveraged a blend of time-series and regression techniques to develop these models.
Home value prediction: Enhanced existing Auto Valuation Models by incorporating real estate data, including comparable sales, property characteristics, and price trends, to more accurately predict property values.
High risk postcodes: Utilized Generalized Linear Models to analyze market trends and accurately predict postcodes in Australia that would experience a decline in property prices.
Propensity to List: Engineered Smart List product by harnessing property, market, and consumer data to identify properties likely to be listed for sale in next 3 months. This strategic approach aids real estate agents in maximising appraisal and sales opportunities through targeted conversations.

Churn Modelling: Engineered customer churn models for B2B by reviewing clients' CRM datasets and click-stream databases. Developed features and constructed machine learning and decision tree models to predict churn probability and identify key drivers of customer churn.

Worked for one of the biggest retailer's and also top FMCG suppliers in Australia. I played a pivotal role in engaging with internal stakeholders to gather project requirements.
A/B testing: Designed and implemented an algorithm for clustering retail stores and subsequently selecting test and control stores. The algorithm effectively assesses the impact of various store trials, including layout changes and the addition of new aisles, utilizing significance tests for comprehensive evaluation.
Promotion uplift modelling: Created the Promotion Effectiveness Tool to analyze the impact of various promotional drivers. Utilized a blend of baseline models and uplift models for comprehensive assessment.
Deletion analysis: Engineered a deletion tool employing descriptive analytics to estimate the impact of removing a product from the store. This tool enables FMCG suppliers to strategically defend and retain their products within the store.

Conducted a thorough analysis of the Australian insurance market, offering valuable insights on market positioning to insurance companies.

• Deconstructed insurance pricing models, extracting actionable insights for insurance companies.
• Identified pricing gaps between insurer and competitors in various market segments.