Data Analyst · Automation · AI & ML

Hi, I'm Anju Barai.

I build automation systems, analytics dashboards and workflow tools, and I explore machine learning and AI agents on the side. I like turning messy operational data into something people can actually use, and writing about what I learn along the way.

GitHub Kaggle Medium LinkedIn Website

Featured

My latest project, published as an article, a GitHub repo, and on Kaggle and Hugging Face.

Abstract illustration of an application window being scanned by an AI system

AI agents · Reverse engineering

Reverse Engineering with AI: Hands-on REA Demo

A beginner-friendly look at REA (Reverse Engineer Anything), an open-source MCP tool that lets AI coding agents inspect real software and return evidence instead of guesses. Includes a safe demo app, real REA output and a rebuild-and-verify exercise.

MCPAI agentsElectronNode.js

Article GitHub Kaggle notebook HF dataset

Live demos

Try these right in your browser, hosted on Hugging Face Spaces.

🛒 Retail Analytics Toolkit

Shrinkage, footfall and short-expiry risk on a synthetic retail dataset, with branch filters and interactive charts.

GradioSQLPlotly

Open demo →

🏠 California House Price Predictor

Enter a district's features and compare Linear Regression, a Decision Tree and a Random Forest on test-set metrics.

Gradioscikit-learnRegression

Open demo →

📈 SmartMIS Data Quality

Data Quality Scores, validation checks, cleaning steps and quarantined rows produced by the SmartMIS engine on sample MIS data.

pandaspydanticData quality

Open demo →

AI, automation & analytics

Portfolio projects I'm actively building.

Dot Lab

An experimental always-on AI agent that plans, uses tools, keeps memory, runs multi-step tasks and asks for human approval when needed.

PythonAI agentsTool use

GitHub →Kaggle →

SmartMIS

An MIS automation and analytics platform: data validation, a KPI engine, TAT/SLA tracking, anomaly detection and automated reports.

PythonpandasStreamlitSQLAlchemy

GitHub →Demo →

Retail Analytics Toolkit

A SQL + Python toolkit for retail shrinkage, footfall and short-expiry analysis, built on synthetic data.

SQLPythonRetail analytics

GitHub →Demo →

Risk Analytics Deployment

A project focused on deploying risk analytics solutions using statistical and machine learning techniques.

Machine learningDeployment

GitHub →

Machine learning & deep learning

Hands-on projects across NLP, computer vision and classic ML.

NLP + Computer Vision Project

Tweet sentiment classification with SVM and Word2Vec, plus transfer learning for house image classification.

NLPWord2VecTransfer learning

GitHub →

Ghibli-Style Image Converter

A CycleGAN-based converter that turns photos into Ghibli-inspired artwork, with a Gradio interface designed for Colab GPUs.

CycleGANGradioGenerative AI

GitHub →

Vehicle Classification (CNN)

A CNN-based vehicle classifier built with TensorFlow/Keras.

CNNTensorFlowKeras

GitHub →

Scene Image Classification (CNN)

A CNN that classifies images into Building, Forest and Sea categories.

CNNComputer vision

GitHub →

Employee Wellness Prediction

An end-to-end pipeline covering data cleaning, encoding and model training to predict employee wellness.

Classificationscikit-learn

GitHub →

Regression & Classification Basics

House price prediction with linear regression, and income (>50K) prediction with logistic regression.

Linear regressionLogistic regression

Linear → Logistic → Demo →

Dashboards & BI

Interactive reporting work.

Power BI Analysis Collection

Dashboards for Budget vs Sales, Retail Sales, School Data, Heart Disease, Titanic Survival and more, each built on its own dataset.

Power BIDAXData visualization

GitHub →

Tableau Project

Visual analytics work in Tableau.

Tableau

GitHub →

Toolbox

Data & analysis

SQL, Python, pandas, Excel, Google Sheets

BI & dashboards

Power BI, Looker Studio, Tableau, Streamlit

Automation

Google Apps Script, workflow tools, automated reporting

ML & AI

scikit-learn, TensorFlow/Keras, NLP, computer vision, AI agents, MCP