Ad Targeting and Campaign Automation

Targeted 100K+ customers by building Power Automate pipelines integrated with Python ML workflows and PostgreSQL segmentation queries. Later, launched a Power BI dashboard that automated experimentation and tracking of 10,000+ GPT-personalized emails.

GTM & Growth Analytics - Copilot enabled

Built a multi-tab GTM intelligence MVP for a YC-backed commercial insurance brokerage unifying leads, net revenue, ad spend, policies, and automation activity into actionable KPI views with a data-grounded Live Co-Pilot for NL analytics that answers GTM question.

Automated Retail Email Agent — Order Fulfillment & Customer Inquiries

Designed and developed an end-to-end AI-powered email automation system for a fashion retail store using GPT-4o and Retrieval-Augmented Generation (RAG). The system intelligently classifies customer emails into order requests or product inquiries, processes orders with real-time inventory validation

Inventory Management: Cloud Automated Solution

An SQL powered solution for inventory forecasting, turnover optimization, and transactions analysis, enhancing resource allocation and consumer service for stakeholders.

US Coffee Sales Dashboard: From Data to Strategy

Analyzed regional coffee product sales across the US, structured & aggregated data, designed dynamic charts, and presented data-driven insights to optimize vendor distribution, inventory, and marketing strategies

HRIS - Payroll Analysis & Employee Management

Coda-based HRIS with OpenAI-Copilot to automate comp benchmarking, offer workflows, and compliance checks. Added KPI-based guardrails for market vs internal bands comparison - supporting employee benefits and payroll

TalentLink - Smart Matching, Faster Hiring

Developed a Streamlit-based HR automation tool using AWS Bedrock and GPT-3.5 to refine job descriptions and rank resumes. Achieved 95% matching accuracy, cut HR evaluation time by 80%, and improved hiring precision by 40%. Automated manual tasks and supported structured job data for compensation.

Apexbot: Conversational Customer Support & Intelligence

An opensource Generative AI usecase capable of resolving queries based on data available on a website. The MVP is deployed on https://www.apexon.com, a $2B global technology services firm specializing in accelerating business transformation & delivering human-centric experience.

Affinite.ai

This AI-driven matchmaking recommendation system utilizes advanced machine learning techniques to facilitate meaningful connections. By analyzing user profiles and interactions, the system employs collaborative filtering and neural networks to predict and recommend compatible matches.

FraudTrack: Fraud Detection & Real-time Monitoring

Developed an accurate fraud detection model using K-Means, Z-score, Isolation Forest, & Logistic Regression, trained on 2,000+ txns, analyzing, predicting & classifying frauds based on user activity pattern. Built PowerBI Dashboard for Real-time threat monitoring.

Credit Card Fraud Detection

A simple and fast enough Machine Learning model to detect fraudulent transactions with a 99.9% accuracy using Machine Learning concepts like Random Forest Classifier, Logistic Regression, and Gradient Boosting.

Impact of FAANG Companies on San Francisco Bay Area Housing Prices

This project examines how FAANG companies affect Bay Area housing prices, employing advanced machine learning models to achieve 99.9% accuracy. The analysis highlights the significant economic influence of tech offices on local real estate markets.

GDP VS Life Expectancy Trend in Top Economies

Analytics of GDP’s influence on life expectancy & healthcare infrastructure in major economies, projections of 2030 life-expectancy rates. To achieve the results, raw WHO datasets has been processed, for regression and classification and the insights has been shared with an audience of 80+ people.

Recession Forecasting Model: United States of America

The project developed a 90% accurate model predicting U.S. recessions by analyzing 8 Key indicators from past economic downturns using data pre-processing pipelines and machine learning algorithms. This tool aids in anticipating economic downturns and facilitating proactive measures.