Supriyo Ghosh

Developer
Download CV

Supriyo Ghosh

Developer

Empowering businesses with cutting-edge full-stack development and machine learning expertise. I deliver tailored, scalable solutions that drive innovation, enhance user experience, and accelerate growth for your success.

Collaborated with elite professionals

tap to know more

Trusted by Leading Organisations

I

About Me.

Hi, my name is Supriyo Ghosh. I am a computer science engineer skilled in full-stack development (React, Next.js, PHP, Node.js, MongoDB) and ML (TensorFlow, Scikit-learn). With experience at ISRO and CodeClouds, I've developed cutting-edge systems like an AI-based Mars landing system and a Visitor Management Platform. My projects include a service booking app, Flipkart sales forecasting using ARIMA, event management site, a scalable social media web app etc. I have been recognized as a Cognizant Hackathon finalist and endorsed by ISRO's Dr. Ritu Karidhal. I am also a winner of multiple technical events. I bring technical excellence, innovation, and a proven problem-solving record to every project.

Name

Location

Experience

Full-time

Language

: Supriyo Ghosh

: Kolkata, India

: 1+ years

: Available

: English

Education.

B.TECH in Computer Science Engineering

MCKV Institute of Engineering (MAKAUT) / 2021 - 2025

Grade: 9.67 CGPA (upto 3rd year)

Higher Secondary (12th) - PCMB

Ramakrishna Mission Vidyapith, Purulia / 2019 - 2021

Grade: 89 percentage

Secondary (10th)

Ramakrishna Ashrama Vidyapitha, Panagarh / 2014

Grade: 94% with 98% in Maths

Experience.

Backend Developer Intern

Codeclouds, Kolkata / 2024 Feb - Present

PHP, MySQL, Backend Development. Developed a scalable Visitor Management System backend, featuring OTP-based check-in, admin tools, and real-time performance.

ML Intern

ISRO, Bangalore / 2024 Jan - 2024 Feb

ML, Neural Networks, tensorflow, scikit-learn, unet, computer vision. Developed an AI Mars landing system using deep learning for terrain classification, achieving 86% accuracy with self prepared dataset

Web Development Intern

Oasis Infobyte / 2023 Jan - 2023 Feb

HTML, CSS and JavaScript. Developed and maintained three frontend web applications.

Coding Skills.

92%
C++for DSA
88%
JSfor Dev
85%
Pythonfor ML
82%
Javafor Oops

Tech Stack.

React JS

PHP

Tensorflow

MySQL

Wordpress

Git

Awards.

2024

Cognizant DN Hackathon 2024

Finalist

tap to know more

AI-driven personalized health platform that predicts disease risks, provides tailored recommendations, and enables real-time health monitoring

2024

ISRO

Recommendation

tap to know more

Achieved 'Recommendation Letter' from Ritu Karidhal, the "Rocket Woman of India", for tenacious effort and comprehensive display of AI training techniques at ISRO.

2023

TOI Debate Competition

Zonal Winner

tap to know more

Displayed expertise in communication and logical thinking in a state level debate competition "Speak for India".

2023

Innovative Idea Presentation

2nd Runner Up

tap to know more

Achieved position in the innovative idea presentation competition at annual intercollege Tech-Fest for the groundbreaking concept of "Brain-Controlled Car Using AI."

Services.

BUILDING DIGITAL PRODUCT WITH BETTER EXPERIENCE

My vision is to be a trailblazing force in the world of web design and development and ML recognized for our unwavering commitment to excellence, integrity, and customer satisfaction.

MACHINE LEARNING

Unlock actionable insights with custom AI models and solutions. Let's collaborate to drive innovation and efficiency!

BACKEND DEVELOPMENT

Optimize your systems with scalable, secure backend solutions tailored to your business. Let's build seamless solutions for success!

WEB DEVELOPMENT

Enhance your online presence with custom, responsive websites tailored to your business needs. Let's build something exceptional!

BACKEND DEVELOPMENT

Optimize your systems with scalable, secure backend solutions tailored to your business. Let's build seamless solutions for success!

Portfolio.

OneUtil - Utility Service Booking

OneUtil

Utility Service Booking

tap to know more

ConnectX - Social Media Webapp

ConnectX

Social Media Webapp

tap to know more

Insignia - Event Management Site

Insignia

Event Management Site

tap to know more

MyOnlineMeal - Food Delivery Site

MyOnlineMeal

Food Delivery Site

tap to know more

NewClass App - Brackettup Assignment

NewClass App

Brackettup Assignment

tap to know more

Sales Prediction of a Store

Walmart Sales Prediction

Time Series

tap to know more

Testimonials.

Their words are a testament to our commitment to excellence and our ability to bring their visions to life. Read on to see how we've made a lasting impact on their brands and projects.

I am sure that Mr. Supriyo Ghosh is capable enough to hone his skills in the domain of Al technologies to realize advanced products or services in this field. I would like to offer my highest recommendation for him for his future endeavours.

Smt. Ritu Karishal

Director-MPOG, ISRO

click to show document

He is a student of Bachelor of Technology Computer Science & Engineering course. I am positively impressed by his sincere and diligent nature towards academics. He is dedicated towards perfecting his understanding of subjects, which is a quality I admire as a teacher.

Prof. Avijit Bose

HOD, CSE-MCKVIE

click to show document

During the internship he demonstrated good coding skills with a self-motivated attitude to learn new things. His performance exceeded expectations and was able to complete the project successfully on time. We look forward to achieving more milestones together.

Subhojit Ganguly

Chief Operating Officer,
Codeclouds

click to show document

Blogs.

11 NOV

11 November 2024 | MCKVIE

Multidimensional Medical Image Segmentation: A Deep Learning Based Approach

Medical image segmentation is considered one of the most significant tasks in clinical practice: it allows for accurate diagnosis...

13 FEB

13 February | ISRO

AI-Powered Terrain Assessment for Autonomous Landing

Landing on Mars is a complex and risky endeavour, requiring precise navigation and hazard avoidance amidst a vastly different environment. Our project, "Automated Landing on Mars" tackles this challenge by...

Get In Touch.

TAKE A COFFEE & CHAT WITH ME

Email

supriyoghosh27125@gmail.com

Phone

+91 6296043650

© 2025 | All rights reserved by Supriyo Ghosh

Collaborated with exceptional Talents

Ritu Karidhal

Director - MPOG, ISRO

View my work with her

Ritu Karidhal Srivastava is an Indian scientist and aerospace engineer working in the Indian Space Research Organisation (ISRO). She was a Deputy Operations Director to India's Mars orbital mission, Mangalyaan. She has been referred to as one of the many "Rocket Women" of India.

kaustubh K. Anand

Scientist, ISRO

View my work with him

A brilliant scientist at ISRO's URSC Bangalore, pushing the boundaries of space exploration through innovative research in machine learning and AI. His contributions are invaluable to the advancement of India's space program.

Somnath Bhattacharyya

Director - Learning & Development, Codeclouds

View my work with him

A highly experienced IT professional with a strong background in backend management for web applications. He brings a wealth of industry knowledge and has been a key contributor to Codeclouds's success for many years.

Puspen Lahiri

Professor - CSE, MCKVIE

View my work with him

A highly regarded CSE professor at MCKV Institute of Engineering, with expertise in Machine Learning and Data Structures. He is an active researcher with numerous publications and a proven track record of successful research projects.

Oneutil

This project is a full-stack web application built using the MERN (MongoDB, Express.js, React, Node.js) stack. It provides a platform for users to book various home utility services, including cleaning, plumbing, electrical repairs, and more.

Live site: oneutil.onrender.com

(Note: May take ~2 minutes to load due to free-tier hosting on Render.)

Github: github.com/Supriyo02/OneUtil

Technologies:

Frontend

React, TypeScript, Vite, Tailwind CSS, React Router Dom

Backend

Node.js, Express.js, MongoDB, bcryptjs, Stripe

Testing

Playwright

Key Features:

  • User Authentication: Users can create accounts, log in, and securely manage their profiles using bcryptjs for password hashing and session cookies for authentication.
  • Service Creation and Management: Users can create their own services, providing detailed information such as service type, price, location, category, facilities, and images. They can also edit and manage their existing services.
  • Search and Filtering: The application offers advanced search and filtering capabilities, allowing users to find services based on location, service type, rating, price, and other criteria.
  • Booking and Payment: Users can book a service for a specific date and time. The booking process includes payment integration using Stripe for secure transactions.
  • Booking Management: Both users and service providers can view and manage their booked services, including confirmation details and status.
  • Error Handling: The backend API incorporates robust error handling mechanisms to prevent crashes and provide informative error messages to users.
  • Data Storage: MongoDB is used to store user data, service information, and booking details in a secure and scalable manner.
  • Frontend Development: The frontend is built using React, providing a modern and interactive user interface. Tailwind CSS is used for styling, and React Router Dom for navigation between different pages.
  • Testing: The application undergoes thorough end-to-end testing using Playwright to ensure quality and prevent bugs.
  • Deployment: The final product is deployed on Render for easy access and availability.

ConnectX

This project is a cutting-edge social media platform developed using Nextjs, reactjs, Tailwind CSS, and Supabase.

Live site: connectxt.vercel.app

Note: Request owner once before accessing the site (The backend database may need to be reactivated as it is hosted in free tier of supabase).

Github: github.com/Supriyo02/ConnectBack

Technologies:

Frontend

Next.js, React.js, Tailwind CSS

Backend

Supabase

Key Features:

  • High Performance: The app's Next.js framework and optimized codebase ensure fast loading times and a smooth user experience.
  • Engaging Design: Tailwind CSS provides a clean and visually appealing interface, enhancing user engagement.
  • Robust Backend: Supabase offers a scalable and secure backend, handling data management, authentication, and user privacy efficiently.
  • User Authentication: Users can create accounts, log in, and securely manage their profiles.
  • Post Creation and Management: Users can create, edit, and delete posts, including text, images, and videos.
  • Social Interaction: Users can like, comment, and share posts, as well as follow other users.
  • Post Saving: Users can save posts for later viewing.
  • Search Functionality: Users can search for other users and posts.
  • User Profiles: Each user has a personalized profile page with information and activity feed.

Insignia

Developed an efficient frontend website to deliver event details and to connect with more students of our college for "Insignia", a technical event organised by CSE & CSE-DS of MCKVIE.

Live site: insignia-2k23.netlify.app

Note: Wait for a moment after clicking the link (it may take some time to load as it is not maintained now).

Github: github.com/Supriyo02/Insignia.23

Technologies:

HTML, JavaScript, CSS

Key Features:

  • Student Registration: Users can register for the event using a Google Form integration.
  • Dynamic Timer: A countdown timer displays the remaining time until the registration deadline.
  • Event Alerts: Users can add event details to their Google Calendar for reminders.
  • Event Details: Comprehensive information about each round of the event is provided, including rules, schedule, and judging criteria.

Achievements:

  • Increased User Engagement: The website achieved a 20% increase in user engagement compared to previous event promotion methods.
  • Successful Registrations: Over 80 teams registered for the "Insignia" event through the website.

MyOnlineMeal

Online Food Delivery Site: This project is a mobile-responsive frontend website for online food ordering.

Live site: myyonlinemeal.netlify.app

Note: Wait for a moment after clicking the link (it may take some time to load as it is not maintained now).

Github: github.com/Supriyo02/MyOnlineMeal

Technologies:

HTML, CSS

Key Features:

  • Food Catering: Simulates the ability to inquire about catering services.
  • Instant Food Ordering: Mimics the process of placing an instant food order.
  • Bulk Ordering: Offers a frontend interface for bulk order inquiries.
  • Enquiry Form: Provides a user-friendly form for submitting inquiries
  • Mobile Responsive Design: Ensures the website adapts seamlessly for mobile devices.
  • Note: It showcases various functionalities but does not implement actual ordering or business logic (backend).

NewClass App

Brackettup Assignment: This is a react-native project bootstrapped with expo.

Live site: myyonlinemeal.netlify.app

Note: Wait for a moment after opening the app (it may take some time to load as it is not maintained now).

Github: github.com/Supriyo02/ClassApp

Technologies:

React Native, Expo

Methods:

  • Running in android/ios mobiles (Suggested/Efficient method): Install the 'Expo Go' mobile app from playstore/app store. Launch the app and connect with the same (wifi)network. (In ios devices, give 'local network' access in the expo app settings and turn of vpn services from the device settings) Scan the qr code provided in the development terminal of the running project from the app(or from camera app in ios). Boom! The app starts running on your device. Press 'r' in the running terminal to refresh the app if you encounter any issue. If the issue persists try clearing the cache in expo go mobile app and restart the app.
  • Running in web: While the project is running press 'w' to run in web. Open the localhost link provided in the terminal.
  • Running in android/ios emulator: Press 'a' for android or 'i' for ios devices. The emulator will start and the app will start running in the virtual device. Please ensure that android studio virtual device or a emulator is installed on your pc and is ready to run. Among these all three options, first option, opening the app using your mobile handset is highly recommended. It's also a easy and handy method with a feel of hosted app. Though the app will run in both android/ios device, it is optimized for android devices.

Walmart Sales Prediction

In this project a Time series ML model have been created to predict the sales of a Walmart store, using the previous sales data of the store, consisting of some other exogenous variabled.

Live site: Walmart Sales Prediction

Github: Walmart Sales Prediction

Technologies:

ML, Sarimax, scikit-learn

Description:

  • The model comes with 96% accuracy score and fits well with the actual data. It can forecast the future sells of a store almost accurately with the time. The accuracy of the model increases even more if we can provide it some data of the future. It is a beginner friendly simple model, but beats other complex models in it's accuracy score.
  • SARIMAX model have been used for fiiting the model, as our time series dataset has seasonal component and it also consist of exogenous variables. Some data preprocessing is done before training the model. seasonal_decompose is used for decomposing the dataset and get a insight for any trend, seasonality and residual pattern. I have used auro_arima to estimate the parameters (p,d,q), (P,D,Q,m) and from the time series graph and have visually estimated the probable seasonal duration(m). The exogenous variables are taken in consideration as they affects our target variable, and so the model accuracy has also been increased. The dataset is splitten in test and training dataset and the SARIMAX model is trained on the training dataset. After training the model, predictions are made and they are plotted against the actual test dataset, and they fits well. Lastly the accuracy of the model is calculated.

Multidimensional Medical Image Segmentation: A Deep Learning Based Approach

11 November 2024 | Computer Vision
- by Supriyo Ghosh, Arkaprabha Laha, Arijit Saha, Shubradip Chowdhury, Subhomay Kundu Poddar

Medical image segmentation is considered one of the most significant tasks in clinical practice: it allows for accurate diagnosis, treatment planning, and monitoring of diseases. It has been shown that deep learning models, especially U-Net and its variants, are beneficial in this field. Despite these promising results, challenging issues remain in the fast and accurate segmentation of complex medical images.
Medical image segmentation simply translates into dividing the image into meaningful regions. Accurate segmentation lets anatomical structures be measured precisely. It detects lesions and measures disease progression. A new tool in medical image segmentation has matured lately with deep learning-based models, particularly convolutional neural networks or CNNs.
Recently, U-Net has become popular CNN architecture in medical image segmentation because it can capture not only the high-level semantic information but also the low-level spatial details. The traditional U-Net architecture, however, often fails to segment complex structures in volumetric medical images. To overcome such weaknesses, a variety of enhancements were proposed, such as adding attention mechanisms, residual connections, and multi-scale feature extraction.
We now propose an architecture for a new type of U-Net by improving these aspects. Here, we describe the important components of our architecture as follows:
Dense Inception-Res Block: This block connects dense connections, inception modules and residual learning to extract multi-scale features efficiently.
Attention Gates: These gates selectively refine features across skip connections by focusing on the most important information and reducing redundant feature transmission.
Multi-Scale Fusion: This method combines information across scales, contextualizing more insightful and finely accurate segmentation.
These components are going to achieve better performance of the architecture in the complex medical images, particularly MRI scans. The proposed approach is going to advance the field of medical image analysis for clinical decision-making purposes.

AI-Powered Terrain Assessment for Autonomous Landing

13 February | Machine Learning
- by Supriyo Ghosh, Arijit Saha

Landing on Mars is a complex and risky endeavour, requiring precise navigation and hazard avoidance amidst a vastly different environment. Our project, "Automated Landing on Mars" tackles this challenge by leveraging the power of Deep Learning to analyse the Martian surface and identify safe landing zones for future missions.
We embark on this journey by utilizing high-resolution images captured by the MRO CTX. Extracting valuable insights from the labeled data, we transform the labeled information into grayscale masks. These masks serve as a critical intermediate step, translating the complex visual information into a format readily digestible by our Deep Learning model.
At the heart of our approach lies a strong neural network model. we train our model to identify safe landing zones with remarkable accuracy. Our rigorous training efforts yield an impressive 77% accuracy, signifying the effectiveness of our chosen approach. Beyond the training data, the true test lies in applying our model to unlabeled images, replicating the real-world scenario encountered by future missions. Pushing the boundaries, we test our model on unseen Martian terrain, and the results are encouraging, surpassing our satisfaction levels.
This project represents a significant step towards enabling safe and efficient automated landing on Mars. By harnessing the power of Deep Learning and meticulously crafted training data, we pave the way for future missions to confidently explore and land on the Red Planet. Prepare to be captivated as we delve deeper into the technical specifics and fascinating results of our endeavors.