Shwetabh Biswas
Artificial Intelligence | Machine Learning | NLP | CV | Data Science | Life Long Learner
Welcome to my portfolio website! This repository showcases my work, projects, and skills. Feel free to explore and learn more about what I do.
Table of Contents
This is the link to my resume : Resume
About Me
Hello! I’m Shwetabh Biswas, a recent graduate from IIT Jodhpur with a passion for exploring the exciting realm of Artificial Intelligence. My academic journey in physics ignited my curiosity for ML, leading me to delve into various aspects of this fascinating field.
Here are five key points that define my journey:
AI Enthusiast: My interest in ML sparked during my master’s program. I dedicated myself to understanding the fundamentals by completing essential courses like CS231n and CS224n. To deepen my knowledge, I embarked on side projects, including creating a transformer architecture from scratch for neural machine translation and training a variational autoencoder for image generation.
Project Enthusiast: I have a penchant for turning ideas into reality. I’ve worked on diverse projects, such as developing chatbots and AI-assistants capable of facial recognition using MTCNN and FaceNet architectures. These projects are not just about technology but also about creating meaningful interactions. One standout project is Llama-2, a conversational agent equipped with text-to-speech and speech-to-text capabilities, seamlessly handling tasks like answering questions and reading emails.
Seeking Opportunities: I am currently on the lookout for project opportunities or internships that will not only deepen my understanding of NLP and CV but also serve as a springboard for my academic career in NLP, CV, and related fields. I am eager to contribute my skills and knowledge to innovative projects.
Location: I am based in Bangalore. If you have any exciting opportunities aligning with my interests, please don’t hesitate to reach out. I am enthusiastic about collaborating on projects that push the boundaries of NLP/CV and contribute to the evolving landscape of artificial intelligence.
Diverse Skill Set: My experiences extend beyond coursework and projects. I have honed a range of skills, from developing DNN architectures to deploying them on a server using Flask/REST-API. My commitment to continuous learning and hands-on experience sets me apart in the dynamic field of NLP.
I am excited about the possibilities that lie ahead and welcome any opportunities for collaboration, learning, and growth. Feel free to connect with me if you have a project in mind or if you simply want to discuss the fascinating world of AI!
Education
- Master of Science , Physics Indian Institute of Technology, Jodhpur May 2021 - June 2023
- Bachelor of science, Physics Center For Basic Sciences, Raipur, CG May 2018 - June 2021
Certifications
- Python For Data Science And Machine Learning Bootcamp - Udemy
- CS231n - Convolutional Neural Networks For Visual Recognition - Stanford
- CS223n - Natural Language Processing - Stanford
- Deployment Of Machine Learning Models- Udemy
- Data Analytics Certification - WSCube Tech
- Data Driven Astronomy - Coursera
Blogs
- I recently wrote a blog on Understanding Multimodal Large Language Models (MLLMs). click here to read it.
- A Medium version of the same blog is also available for better readibility and visuals. click here to read it.
Projects
Here are some of the key projects I have worked on:
Person Detection And Cropping : For a detailed overview, click here
- Integrated YOLOv8 and FaceNet for embedding generation and cropping out people and faces.
- Deployed on Railway app using Dockerization and FAST-API.
Image Reconstruction Using a VAE : For a detailed overview, click here
- Implementing a CNN-based VAE architecture from scratch
- Learned and implemented the parametrization trick to back-propagate the losses successfully.
- Explored the architecture and performance over a variety of CNN layers(128, 256, 512).
AI-powered Assistant : For a detailed overview, click here
- Conversational Abilities: Leveraged the advanced language model LLM Llama-2 to power the conversational aspect of the AI assistant and also, used quantization to reduce the modelsize from 20 GB to 6 GB.
- Face Recognition using Facenet and Object Detection using YOLO: Learned the architectures of YOLO and FaceNet.
- Integration of various models and APIsto read emails and check calendars.
Skills
- Programming Languages: Python
- Web Development: Flask, FastAPI
- Database: MySQL
- Tools & Technologies: Git, Docker,
Contact
Feel free to reach out to me:
- Email: send me an email
- LinkedIn: Let’s connect on LinkedIn
- GitHub: See my projects
- Instagram: have a look at my social life
I am open to collaboration and new opportunities. Let’s connect!