About Me

🧠 Pursuing Data Science because apparently, analyzing all that data is more fun than having a social life. Can't wait to unlock the secrets of the universe hidden within those spreadsheets and CSV files.

📚 Always on a quest to learn and get exposed to the industry, because who needs free time and hobbies when you can drown yourself in code and algorithms? Ready to absorb all the knowledge and become the ultimate data guru.

🏋️‍♂️ When I'm not crunching numbers, you'll find me doing some serious lifting at the data gym. Yes, that's right, I'm a data scientist by day and a gym enthusiast by... well, whenever I have the energy. Turns out, deadlifts and debugging have a lot in common – both require strong determination and a solid grip!

Concepts

Data Science Data Structures & Algorithms Generative AI Prompt Engineering Automation Object-Oriented Programming Machine Learning Deep Learning Databases Image Analytics Software Engineering Full Stack Development Operating Systems Statistics Mathematics

Tools/Frameworks

Python R HTML5 - CSS JavaScript GithubPages PythonAnywhere Render Postgres MySQL MongoDB Keras Plotly PyTorch TensorFlow GitHub Git Power Bi Postman

Projects

(Click/Hover to see more details)

Real-time chatbot with web scraping for dynamic responses.

Live Dynamic Chatbot

● Developed a chatbot that scrapes web content and generates context-based answers in real-time.

● Created a local vector database (ChromaDB) for efficient storage and retrieval of embeddings.

● Used Langchain and OpenAI libraries to convert data into embeddings for accurate responses.

Image classification pipeline for exercise types using CNN.

Exercise Classification

● Originated a comprehensive ML pipeline for classifying images of hand-stands, push-ups, and crunches

● Utilized TensorFlow, Keras, YOLO v5 and CNN algorithm for model training

Interactive digital album for organizing and viewing photos.

Photo Album

• Created a web application using Flask to display a photo album organized in a grid layout with hover effects

• Incorporated HTML templates to render static content and CSS for styling and responsiveness

• Implemented routes for navigating through the album and individual photos

Automated signature verification system with auto-cropping.

Signature Verification

● Set up a web-based system that verifies the signature of users based on previously collected signatures

● Developed a mechanism for auto-cropping and mapping the signatures of individuals during the processing stage

Estimating Sonic Layer Depth (oceanography)

Estimating Sonic Layer Depth from Satellite Observations

● Developed a model to estimate Sonic Layer Depth in the Indian Ocean using satellite observations and Argo float data, focusing on achieving high accuracy.

● Conducted data wrangling and analysis of both in-situ and satellite data using Python and CDO tools.

● After training and tuning multiple models, the XGBoost Regressor emerged as the most accurate for the task.

Hey, I'm Krish Agarwal