About Me โน๏ธ
I am currently pursuing an MSc in Data Science as part of a 5-year integrated program at PSG College of Technology, currently in my 5th year. My primary area of interest lies in data science, particularly in supervised and unsupervised learning, along with LLM.
Projects ๐ ๏ธ
SecureSwipe ๐ณ
- Description: Developed a fraudulent credit card detection website using Streamlit.
- Technologies Used: Streamlit, Python, Machine Learning
- Link: SecureSwipe Repository
ExchangeInsight ๐ฑ
- Description: Utilized Python to predict foreign currency exchange rates.
- Technologies Used: Python, Machine Learning, freecurrencyapi.com
- Link: ExchangeInsight Repository
DeepDoom-Learning-to-Survive ๐ฎ
- Description: DeepDoom is a deep reinforcement learning project built on top of ViZDoom, a Doom-based RL platform. This project explores the training and evaluation of AI agents using PPO, A2C, and DQN algorithms in complex FPS game scenarios.
- Technologies Used: Python , PPO , A2C, DQN
- Link: Deep Doom Repository
Multimodal-Sign-Braille-Translator ๐
- Description: Built a Flask web app that takes input from speech, text, or images and converts it to Indian Sign Language (ISL) and Braille.We used Whisper, Google Speech API, and EasyOCR for transcription, along with language detection, translation, and grammar correction for better accessibility.
- Technologies Used: Python ,Flask, OpenAI Whisper ,Google Speech Recognition API, EasyOCR, Langdetect, T5 (vennify/t5-base-grammar-correction), MarianMT (Helsinki-NLP), OpenCV ,MoviePy ,PIL
- Link: Multimodal_Sign_Braille_Translator Repository
ArtisticAlchemy ๐จ
- Description: Implemented an Image Diffusion Model leveraging a pretrained stable diffusion model, integrating Variational Autoencoders (VAEs) and UNet models for advanced image generation.
- Technologies Used: Python, Diffusion Models, PyTorch, Transformers
- Link: ArtisticAlchemy Repository
CallWizard โ๏ธ
- Description: Developed a comprehensive phone recommendation system utilizing SQL for database connectivity, BeautifulSoup for web scraping GSMArena, HTML/CSS for the user interface, and Power BI for data visualization.
- Technologies Used: SQL, BeautifulSoup, HTML/CSS, Power BI
- Link: CallWizard Repository
SecuHash ๐
- Description: Implemented a Merkle Trees project in C++ using a simple hash function to ensure data integrity and security in technologies like Blockchain.
- Technologies Used: C++, Merkle Trees, Blockchain
- Link: SecuHash Repository
Skills ๐ง
- Programming Languages: Python, C++
- Data Science: Machine Learning, Data Visualization
- Tools: Streamlit, Power BI
- Other Skills: SQL, Web Scraping
Contact Me ๐ง
- Email: kanishkakannan484@gmail.com
- LinkedIn: Kanishka Kannan on LinkedIn
Let's Connect ๐ Feel free to reach out to me for collaboration opportunities or to discuss anything related to data science and machine learning!