Available for Remote Roles

Data Scientist
& ML Engineer

I build production-grade ML and Agentic AI systems — not just notebooks. Shipped a live AutoML platform powered by GPT-4o, SHAP, and full MLOps pipelines. Available EU/UK overlap daily — I deliver while you sleep.

0.83
R² Score
4+
Live Projects
11
Module Files
5
ML Models
What I Know
Technical Skills
🐍

Languages & Data

Python SQL Pandas NumPy EDA
🤖

Machine Learning

XGBoost LightGBM CatBoost Random Forest Regression Classification SMOTE
🧠

Explainable AI

SHAP Feature Importance Model Interpretability Business Insights

MLOps & Deployment

Streamlit Model Registry Experiment Tracking Batch Pipelines Leakage Detection
🔮

AI & GenAI

GPT-4o API OpenAI SDK Prompt Engineering LLM Integration RAG (Conceptual)
🛠️

Tools

Git GitHub Jupyter MySQL Kaggle
What I've Built
Live Projects
🚀
LIVE

DataPilot AI — Agentic AutoML

Upload any CSV → 5 ML models trained automatically → GPT-4o explains results in plain English → Chat with your data → Download PDF report. Real system, not a notebook.

Result CatBoost R² 0.83 on 10K rows
Stack GPT-4o · SHAP · XGBoost · LightGBM · CatBoost
Agentic AI MLOps SHAP GPT-4o Streamlit
🛒
LIVE

Walmart Sales Forecasting

Regression-based forecasting using retail transaction data, economic indicators, and holiday features. Cyclic date encoding, lag features, LightGBM + XGBoost ensemble.

Result LightGBM ensemble — strong generalization
Stack LightGBM · XGBoost · Feature Engineering
Time-Series Forecasting LightGBM XGBoost
📉
LIVE

Customer Churn Prediction

End-to-end churn prediction using Logistic Regression with SMOTE for class imbalance. Engineered tenure_monthly_interaction feature. Actionable retention recommendations.

Result SMOTE + Logistic Regression — balanced classes
Stack Scikit-learn · SMOTE · Feature Engineering
Classification SMOTE Churn Business Insights
My Approach
How I Think
01

Business First

Every model exists to solve a business problem. I start with "what decision does this enable?" before touching any code.

02

Data Before Models

Garbage in, garbage out. I spend 60% of time on EDA, feature engineering, and leakage detection before training anything.

03

Explainability is Non-Negotiable

A model no one understands is a model no one uses. SHAP + plain English explanations are built into everything I build.

04

Ship It Live

Every project I build is deployed and accessible. A model sitting in a notebook has zero real-world value.

05

Modular by Default

Each file has one job. My DataPilot has 11 modular files — testable, maintainable, and production-ready from day one.

06

AI as Translator

GPT-4o doesn't replace ML — it translates model math into business language. That's the real value of Agentic AI.

Get In Touch
Let's Build
Something Real

I'm a Data Science engineer from India actively seeking remote DS/ML/AI roles globally. Available 5-6hrs EU/UK overlap daily. Fully async-capable.

Open to Remote Work — Data Scientist · ML Engineer · AI Engineer · Available EU/UK overlap · Fully async-capable · I deliver while you sleep 🚀