Hi, I'm Atul 👋
I'm an AI and data science enthusiast pursuing a B.Tech at IIT Patna. My passion lies in solving real-world problems through machine learning and intelligent systems and participating in research activities. Additionally, I'm a skilled web developer proficient in HTML, CSS, JavaScript, Angular, and MySQL, dedicated to creating seamless and impactful applications.
Panipat, India
About me
Curious about me? Here you have it:

I'm Atul Kumar, a dedicated and enthusiastic AI and data science student at the Indian Institute of Technology (IIT) Patna. My journey in artificial intelligence began in 2021. Since then, I've been passionately diving into the complexities of machine learning, data analysis, and intelligent systems. My academic endeavors and hands-on project experiences have fueled my enthusiasm for creating innovative solutions that blend technical precision and impactful results.

In addition to my AI and data science expertise, I have a solid foundation in web development. My proficiency in HTML, CSS, JavaScript, and Angular allows me to create visually appealing and user-friendly web interfaces. Combining my web development skills with my knowledge of backend technologies like FastAPI and MySQL, I can develop full-stack applications that are both functional and efficient. This skill set enables me to bridge the gap between front-end user experience and backend data processing, creating seamless and impactful digital products.

I am very much a progressive thinker and enjoy working on products end to end, from ideation all the way to development.

Finally, some quick bits about me.

  • Avid learner

  • Project Enthusiast

  • Innovative researcher

  • Skilled web developer

Skills
The skills, tools and technologies I am really good at:
C++
Python
Javascript
Typescript
Angular
FastAPI
Docker
MySQL
Sass/Scss
PyTorch
Transformers
Ollama
Scikit-Learn
Git
Github
Skills
The skills, tools and technologies I am really good at:
C++
Python
Javascript
Typescript
Angular
FastAPI
Docker
MySQL
Sass/Scss
PyTorch
Scikit-Learn
Git
Github
Transformers
Ollama
Experience
Here is a quick summary of my most recent experiences:
Prism Intern @Samsung R&D Institute India
  • Developed a lightweight, on-device inference model for automated thumbnail generation from movies.
  • Designed an interactive module to regenerate thumbnails based on user feedback.
  • Optimized algorithms for real-time usability and scalability in video analysis.
June 2024 - March 2025
Developer Intern @Fastcurve
  • Developed and implemented AI-based Face Recognition Technology for the Enkept product.
  • Built free tools in Angular for the Yukt product, improving customer efficiency in day-to-day operations.
  • Worked with teams to add new technologies to the company’s main products, providing solutions.
June 2024 - September 2024
UG Researcher @IIT Patna
  • Utilized efficient knowledge distillation methods to improve medical image analysis accuracy by 2-6%.
  • Published research papers on medical image classification accepted at Knowledge-Based Systems and MICCAI 2024.
  • Currently working on optimizing federated learning frameworks for improved performance and scalability.
June 2023 - April 2025
Experience
Here is a quick summary of my most recent experiences:
June 2024 - March 2025
Prism Intern @Samsung R&D Institute India
  • Developed a lightweight, on-device inference model for automated thumbnail generation from movies.
  • Designed an interactive module to regenerate thumbnails based on user feedback.
  • Optimized algorithms for real-time usability and scalability in video analysis.
June 2024 - September 2024
Developer Intern @Fastcurve
  • Developed and implemented AI-based Face Recognition Technology for the Enkept product.
  • Built free tools in Angular for the Yukt product, improving customer efficiency in day-to-day operations.
  • Worked with teams to add new technologies to the company’s main products, providing solutions.
June 2023 - April 2025
UG Researcher @IIT Patna
  • Utilized efficient knowledge distillation methods to improve medical image analysis accuracy by 2-6%.
  • Published research papers on medical image classification accepted at Knowledge-Based Systems and MICCAI 2024.
  • Currently working on optimizing federated learning frameworks for improved performance and scalability.
Projects
Some of the noteworthy projects I have built:
Context Aware Recipe Generator
The project implements an AI-powered recipe transformation tool designed to customize recipes based on individual dietary preferences, skill levels, and serving sizes. It leverages advanced AI models to suggest ingredient substitutions and provide personalized cooking instructions, ensuring the dish fits specific needs like vegan, gluten-free, or low-carb diets.
FastAPI
LangChain
Ollama LLM
ChromaDB
PyTorch
Angular
Federated Learning
The project implements sentiment analysis using neural networks in a federated setting with PyTorch. We utilized the Sentiment140 and Large Movie Review datasets, along with pre-trained word vectors from GoogleNews-vectors-negative300, to enhance training efficiency. For model selection, we utilize the bidirectional LSTM architecture.
Python
PyTorch
Scikit-Learn
NLTK
Pandas
eStore
The project is an eStore where users can browse and search for items, view detailed product information, and add items to their cart. Users can place orders and create accounts to manage their purchases, track order history, and update personal information. The platform is designed for a seamless shopping experience, integrating essential e-commerce functionalities.
Angular
SCSS
Bootstrap
Node.js
Express.js
MySQL
Projects
Some of the noteworthy projects I have built:
Context Aware Recipe Generator
The project implements an AI-powered recipe transformation tool designed to customize recipes based on individual dietary preferences, skill levels, and serving sizes. It leverages advanced AI models to suggest ingredient substitutions and provide personalized cooking instructions, ensuring the dish fits specific needs like vegan, gluten-free, or low-carb diets.
FastAPI
LangChain
Ollama LLM
ChromaDB
PyTorch
Angular
Federated Learning
The project implements sentiment analysis using neural networks in a federated setting with PyTorch. We utilized the Sentiment140 and Large Movie Review datasets, along with pre-trained word vectors from GoogleNews-vectors-negative300, to enhance training efficiency. For model selection, we utilize the bidirectional LSTM architecture.
Python
PyTorch
Scikit-Learn
NLTK
Pandas
eStore
The project is an eStore where users can browse and search for items, view detailed product information, and add items to their cart. Users can place orders and create accounts to manage their purchases, track order history, and update personal information. The platform is designed for a seamless shopping experience, integrating essential e-commerce functionalities.
Angular
SCSS
Bootstrap
Node.js
Express.js
MySQL
Certificates
Proof of My Accomplishments and Expertise:
Inter IIT Tech Championship 2023
Participated in the Adobe Simulation Challenge, addressing two key tasks: Behavior Simulation (predicting tweet engagement based on content) and Content Simulation (generating tweet text from metadata). This challenge honed my skills in data analysis and content creation, aiding marketers in optimizing social media strategies to achieve desired KPIs.
Angular Intermediate Certificate
Earned the HackerRank Angular Intermediate Certificate, demonstrating proficiency in building dynamic web applications using Angular. This certification validated my skills in Angular concepts, including components, services, directives, and routing, showcasing my ability to develop and maintain robust front-end applications.
Deep Learning Specialization
Completed the Coursera Deep Learning Specialization, mastering key concepts and techniques in deep learning. This certification covered neural networks, convolutional networks, sequence models, and deep learning applications, demonstrating my ability to build and apply deep learning models to real-world problems.
Data Analysis with Python - IBM
Completed the Coursera Data Analysis with Python course, gaining proficiency in data manipulation, visualization, and analysis using Python. This certification covered libraries such as Pandas, NumPy, and Matplotlib, showcasing my ability to analyze data and derive meaningful insights effectively.