Premium Marketing Services | Pune, India 📅 Aug 2025 – Feb 2026 | ⏱️ 6 Months | 🏬 Marketing Department

• 📊 Analyzed marketing data to drive business decision-making
• 📈 Built data visualizations and reports for the marketing team
• 🔍 Applied end-to-end analytics in a real professional environment
• 💡 Delivered actionable insights to support campaign strategies

Python • MySQL • Power BI • ML • Deep Learning • Agentic AI

My flagship capstone — a 12-layer end-to-end Data Science platform built on 100% original Indian e-commerce data.

What's inside:

• 🗄️ MySQL Star Schema — 14,829 records
• 📊 EDA + RFM — 9 customer segments
• 🚨 Anomaly Detection — 34 frauds caught
• 🤖 ML + LSTM + Churn Prediction
• 📊 Power BI Dark Dashboard — DAX
• ⚡ ARTHA Agentic AI — LLaMA 3.3 70B

Python MySQL Power BI TensorFlow Groq Streamlit

🚀 Live App | 📂 GitHub

Trend Recognition and Automated Cause Explanation Python • Streamlit • Groq LLaMA 3.3 70B • Scikit-learn • Plotly

Upload any business CSV or Excel file — TRACE automatically detects anomalies, traces root causes via correlation analysis, and generates plain English diagnostic reports. Any domain. Under 2 minutes.

What's inside:

• 🔬 9 Anomaly Detection Methods — Z-Score, IQR, Isolation Forest, Level Shift & more
• 🔗 Correlation Engine — Lag detection, Pareto ranking, contribution %
• 🤖 LLaMA 3.3 70B — AI-narrated root cause reports via Groq API
• 📊 Plotly Dark Dashboard — KPI cards, heatmap, anomaly timeline
• 🏷️ Confidence Tags — [ VERIFIED ] [ CONFIRMED ] [ PROBABLE ]
• 🌐 Domain Agnostic — Sales, HR, Finance, Logistics, Operations

Skills: Python Streamlit Pandas NumPy SciPy Scikit-learn Plotly Groq API LLaMA 3.3 70B RapidFuzz Statistical Analysis NLP Anomaly Detection Correlation Analysis

🚀 Live App | 📂 GitHub

Excel • MySQL • Python • Power BI • K-Means • Forecasting

4 tools. 4 phases. 785ドル.58M revenue story — zero shortcuts.

Most analysts get clean data. I started with 4,569 rows of corrupted BMW records.

The pipeline:

• 📋 Excel — 29-column audit, 800+ bad records flagged
• 🗄️ MySQL — 8 business queries, CTEs + Window Functions
• 🐍 Python — 13 M5 Black visuals, K-Means (K=4), 193ドル.5M forecast
• 📊 Power BI — 4-page BMW M5 Black dashboard, 13 DAX measures

Python MySQL Power BI DAX Scikit-learn Plotly Excel

📂 GitHub

Python • Groq LLaMA3 • Agentic AI • Voice Enabled

AI agent that autonomously analyzes any CSV/Excel dataset — in Hindi & English.

Highlights:

• 🧠 Groq LLaMA3 — Transformer Agentic AI
• 🎤 Voice Input & Output — Hindi & English
• 📊 Auto Visualizations — 4 chart types
• 💡 Autonomous business insights

Python Groq Streamlit SpeechRecognition gTTS

🚀 Live App | 📂 GitHub

Python • NLP • Machine Learning • Deep Learning

1,459 real reviews from Zomato, Swiggy & Blinkit — 7 models tested.

Highlights:

• 🤖 5 ML + 2 Deep Learning models
• 🏆 Random Forest — 91.10% Accuracy
• 🌐 Live Streamlit App deployed
• 💡 Actionable business insights

Python NLTK TensorFlow Streamlit VADER

🚀 Live App | 📂 GitHub

Python • Scikit-Learn • Random Forest • Forage

Real-world BCG X simulation — challenged the price sensitivity hypothesis with ML.

Highlights:

• 👥 14,606 customers | 9.7% churn rate
• 🤖 Random Forest — ROC-AUC: 0.71
• 💡 Proved price ≠ primary churn driver
• ✅ Certificate — BCG X x Forage

Python Pandas Scikit-Learn Matplotlib Seaborn

📂 GitHub

Python • SciPy • Pandas • Forage

Real retail analytics — 245K+ transactions, uplift testing.

Highlights:

• 🛍️ 245K+ transactions | 71K customers | 1ドル.8M revenue
• 📊 Customer segmentation & uplift testing
• 🔬 Pearson Correlation for control store matching
• ✅ Certificate — Quantium x Forage

Python Pandas SciPy Matplotlib Seaborn

📂 GitHub

Power BI • DAX • Business Intelligence

1ドル.2M+ in sales across 10 outlets — interactive Power BI dashboard.

Highlights:

• 📊 KPIs — Sales, Avg Sales, Items & Ratings
• 🗺️ Sales by Outlet Location (Tier 1/2/3)
• 📈 Trend analysis 2012–2022
• 🔄 Interactive slicers & DAX measures

Power BI DAX Data Visualization Business Intelligence

📂 GitHub

Python • Machine Learning • Predictive Analytics

3 ML models compared on a self-created synthetic pricing dataset.

Highlights:

• 🤖 Linear Regression, Decision Tree, Random Forest
• 🏆 Random Forest — R2: 0.97
• 📉 Residual & error analysis
• 📊 Feature importance breakdown

Python Scikit-Learn Matplotlib Pandas

📂 GitHub

MySQL • Data Analysis • Business Queries

Full e-commerce database built from scratch — 7 real-world business queries solved.

Highlights:

• 🗄️ Designed full schema from scratch
• ❓ 7 business queries answered
• 📋 JOINs | CTEs | Window Functions | CASE WHEN

MySQL MySQL Workbench SQL Data Analysis

📂 GitHub

Python • Google Gemma 3 • Streamlit • Google AI Studio

Built for the Gemma Impact Hackathon 2025 — an AI health companion for 700M+ rural Indians who lack access to doctors.

What's inside:

• 🤖 Google Gemma 3 (gemma-3-4b-it) — Core AI engine
• 🌐 Bilingual — Hindi & English support
• 🚨 Emergency Detection — Auto-alerts with 108/104 helplines
• 💊 6 Quick Symptom Buttons — Accessible for elders
• 🎨 Dark medical green theme — Streamlit Cloud deployed

Python Google Gemma 3 Google AI Studio Streamlit CSS

🚀 Live App | 📂 GitHub

Always building. Always learning.

Currently working on:

• 🧠 ANTARVANI — AI Emotional Support App
• 📊 HR Analytics Dashboard (Power BI)
• 🌍 ATLAS — Conflict-Driven Commodity Dashboard

⭐ Watch this space!