Vishal Ipar

Python And Machine Learning Developer

[Vishal Ipar @IparVishal]

Traditional resume screening relies on keyword matching.

My system uses Sentence Transformers to compare the meaning of a resume with the job description, not just identical words.

#MachineLearning #Python #NLP #SentenceTransformers #Django #AI #BuildInPublic

[Vishal Ipar @IparVishal]

Hiring teams often spend hours manually reviewing resumes for a single role.

I built an AI-powered Resume Screening System that compares resumes with a job description, sort candidates and helps HR shortlist faster.

#Python #Django #MachineLearning #AI #HRTech #BuildInPublic

[Vishal Ipar @IparVishal]

The architecture of vendor invoice processing :

User Input → Backend → Data Preprocessing → ML Models → Predictions

This modular design makes it easy to scale, maintain, and extend with new AI features.

#Python #Django #MachineLearning #AI #SystemDesign #BuildInPublic

[Vishal Ipar @IparVishal]

The second ML feature in my Vendor Invoice Intelligence System predicts whether an invoice should be approved automatically or flagged for manual review.
It helps reduce manual effort and speeds up invoice processing.

Live : https://t.co/8QgKtlKhxb

[Vishal Ipar @IparVishal]

One feature of my Vendor Invoice Intelligence System is Freight Cost Prediction.

Instead of estimating shipping costs manually, the system predicts the expected freight cost from invoice data.

Demo below

#MachineLearning #Python #Django #AI #DataScience #BuildInPublic

[Vishal Ipar @IparVishal]

Manual invoice review is slow, repetitive, and prone to errors.

I built an AI-powered Vendor Invoice Intelligence System that predicts freight costs, flags invoices for manual approval, and provides business insights—all from a simple web dashboard.

[Vishal Ipar @IparVishal]

Building Construction Workers Safety Detection — Custom YOLO Model

A custom object detection system to identify safety helmets (hats) and jackets for construction workers.

Tech Stack : Python, YOLO, OpenCV, LabelImg, Gradio, NumPy

Live : https://t.co/NPOBp3c3yS

[Vishal Ipar @IparVishal]

Customer Churn Prediction
- A project that combines Deep Learning, Artificial Neural Networks, and a Streamlit UI!
- App that predicts whether a customer is likely to churn, based on their input data.

Website Link : https://t.co/6oYvHbjyZP
Source code : https://t.co/9pcdnlMk7r

[Vishal Ipar @IparVishal]

NLP- Natural Language Processing :
Tokenization, Stemming, Lemmatization, POS removal, BoW, TF-IDF, Word2Vec, and AvgWord2Vec
These concepts are the foundation of many modern NLP applications — from chatbots to search engines.

Source code : https://t.co/xQSEoUJytn

[Vishal Ipar @IparVishal]

A Student Performance Prediction system!
This project predicts a student’s Math Score based on the following input features:
Gender
Race/Ethnicity
Parental level of education
Lunch type
Test preparation course
Writing score
Reading score

Source code : https://t.co/uxqgZ3WSiS

[Vishal Ipar @IparVishal]

Clustering Techniques: K-Means, Hierarchical, and DBSCAN :-
It is an unsupervised learning method that groups data points on similarity.
Clustering is a great way to uncover hidden structures in data, from customer segmentation to anomaly detection.

Code: https://t.co/fOFYFBmVmA

[Vishal Ipar @IparVishal]

Principal Component Analysis (PCA) in Machine Learning :-
A dimensionality reduction technique that transforms a large set of variables into a smaller set of principal components, while still retaining most of the variation in the data.

Code : https://t.co/D3KkcdFNcy

[Vishal Ipar @IparVishal]

A diabetes prediction project using decision trees, focusing on two crucial aspects to avoid overfitting and boost performance:
1. Pre-pruning
2. Hyperparameter tuning

Use case -
Data: Patient health metrics (glucose level, BMI, blood pressure).

Code : https://t.co/ztLqbTrM08

[Vishal Ipar @IparVishal]

K-Nearest Neighbors :-
It is a simple yet powerful machine learning algorithm used for both classification and regression. Instead of building a complex model, KNN looks at “K” nearest data points in the training set to make predictions.

Code : https://t.co/piuI8D0swy

[Vishal Ipar @IparVishal]

The Naive Bayes Algorithm in Machine Learning :-
The Naive Bayes algorithm— a simple yet powerful approach in machine learning that’s especially popular for tasks like spam detection, sentiment analysis, and text classification.

Code : https://t.co/OxEn347cZZ

[Vishal Ipar @IparVishal]

Support Vector Machine :-
SVM is a powerful supervised learning algorithm that finds the optimal boundary to separate data into different classes. It focuses on the data points closest to the decision boundary called support vectors.

Code : https://t.co/nXlIl4kkoT

[Vishal Ipar @IparVishal]

Logistic Regression in Machine Learning :-
Logistic Regression is a fundamental supervised learning algorithm used for binary classification problems. Unlike Linear Regression, which predicts continuous values.

Code : https://t.co/eeoXMmjngO

[Vishal Ipar @IparVishal]

Ridge, Lasso and Elastic Net Regression :
Overfitting is a common challenge where the model learns noise than real patterns. Regularization techniques like Ridge, Lasso, and Elastic Net address this by penalizing large coefficients.

Code : https://t.co/ckm9tQHg0N

[Vishal Ipar @IparVishal]

Machine Learning & Data Analysis on E-Commerce Dataset :-
A data analysis project using Machine Learning to gain insights from an e-commerce dataset. This project involved exploring key business metrics to understand sales and profit trends.

Code : https://t.co/J1qRQiLImF

[Vishal Ipar @IparVishal]

Polynomial Regression :-
It captures nonlinear relationships between independent and dependent variables.

Pipeline Concept :
This function creates a polynomial regression model with a specified degree.

Code : https://t.co/8hz4KkOxTH