Narendra Saraf

|

About Me

I'm a software engineer passionate about building intelligent systems. Currently pursuing my degree in Computer Engineering, I spend most of my time working with ML models, building web applications, and solving complex problems through code.

My focus is on creating practical solutions—from predictive models for IPO analysis to full-stack hospital management systems. I enjoy the entire development cycle: understanding the problem, designing the architecture, writing clean code, and deploying production-ready systems.

Profile Picture

Technical Skills

Programming Languages

Python
C++
 Java 
 JavaScript 
  R  
  C  
 SQL 

Data Science & Analytics

Numpy
Pandas
Jupyter
seaborn
Matplotlib
Feature Engineering
Data Cleaning & Preprocessing

AI & Machine Learning

Mathematics
Statistics
PyTorch
TensorFlow
OpenCV
XGBoost
DoWhy
LightBGM
Scikit-Learn
HuggingFace
Fine-Tuning

Web Stack & Backend

React
Express.js
Node.js
Vue.js
FastAPI
Flask
RESTful API
Vue.js
Docker
Kubernetes
AWS

Databases

PostgreSQL
MySQL
SQLite
MongoDB
Redis

Developer Tools & Software

Git
GitHub
GitHub Actions
VS Code
Jupyter Notebook
Virtual Environments
Postman

UI/UX Design & Visualization

Figma
Canva
Wireframes
Prototyping
PowerBI

Projects

Causal Analysis & Bias Mitigation of LLMs

Built a system to identify and reduce bias in language models using causal inference techniques. Instead of just measuring correlation, I analyzed how sensitive attributes actually influence outputs and implemented interventions to reduce unfair predictions. The challenge was maintaining model performance while improving fairness—achieved measurable bias reduction with minimal accuracy trade-off.

HuggingFace Transformers   DoWhy   Matplotlib Numpy Pandas  

IPO Success Scoreboard

Created a machine learning system to evaluate IPO potential using historical financial data. Combined XGBoost and LightGBM models with extensive feature engineering to predict success probability. Added SHAP analysis to make predictions interpretable—investors can see exactly which factors drive the score. Deployed with a web dashboard for real-time evaluation.

Scikit-learn XGBoost LightGBM SHAP Numpy Pandas 

Hospital Management Service Platform

Full-stack web application to manage hospital operations—patient registration, appointments, medical records, and billing. Built with Vue.js frontend and Flask backend, focusing on clean UI/UX and reliable data handling. The goal was to replace fragmented manual processes with a single unified system that doctors and administrators could actually use.

vue.js flask JavaScript typescript 

Eventify

Event management platform handling everything from registration to virtual meetings. Built with the MERN stack—MongoDB for data, Express/Node for the backend, and React for the interface. Integrated real-time notifications and scheduling features. The hardest part was coordinating multiple services (email, video calls, reminders) into a smooth user experience.

MongoDB Node.js React Express.js 

Learnemon

Gamified coding education platform that makes learning feel like playing a game. Built with Flask and SQLite, featuring lessons, quizzes, XP progression, and a simulated terminal for practice. Added competitive elements like leaderboards and achievements to keep users engaged. The idea was simple: make learning code addictive instead of boring.

flask RESTful API Jinja2 Template SQLite HTML CSS JavaScript 

Certifications

Google Cybersecurity Professional Certificate

Coursera

2023

OCI GenAI Professional

Oracle

2025

OCI Data Science Professional

Oracle

2025

OCI AI Vector Search Professional

Oracle

2025

OCI AI Foundation

Oracle

2025

CCNA: Introduction to Networks

Cisco

2024

CCNA: Introduction to Cybersecurity

Cisco

2024

Six Sigma Green Belt

LinkedIn Learning

2025

AWS Essential

LinkedIn Learning

2025

Extra Curricular Activities

Finance Team

Gandharv VIIT

2024

 

Finance and Operations Team

Corpstrata - EDC VIIT

2024

 

Volunteer

EDC VIIT

2023-2024

Let's Connect

Get in Touch

Looking to collaborate on a project or just want to chat about tech? Feel free to reach out.