Welcome to my GitHub profile! I'm a second-year student at IIT Indore with a deep passion for Machine Learning. I have worked on various projects from multiple domains of machine learning like LLMs, NLP, CV and even generative AI.
- Current Education: B.Tech. at IIT Indore
- Interests: Machine Learning, Deep Learning, Generative AI, Quant Finance, Industrial AI, Natural Language Processing, Computer Vision, Data Science, and Business Intelligence
- Current Focus: Enhancing my expertise by building scalable ML systems and contributing to impactful projects.
- Implemented Retrieval Augmented Generation using FAISS Similarity Search on Embeddings of scrapped dataset
- Conducted RL-based ORPO finetuning on DeepSeek-AI-Coder LLM on Github API-scrapped CP dataset
- Utilized LLM with RAG for code generation & explanation solving hard DSA questions & 1200-rated CP questions
– Built an AI-driven system for custom skins generation for Clash of Clans using Stable Diffusion and ControlNets – Implemented finetuning on CompVis-v1-4 model on custom dataset for 200 steps for precise style adaptation – Integrated the finetuned sd model with lllyasviel control-v11p-inpaint for accurate prompt to image generation
– Employed the NinaPro DB5 dataset with 53 diverse gestures from MYO armbands,ensuring thorough preprocessing. – Used various signal processing techniques including rectification, filtering, and smoothing on raw EMG data – Engineered an 86% accurate attention-based model leveraging expansion, attention, and classification layers.
- Programming: Python, C++, SQL
- ML Frameworks: TensorFlow, PyTorch, Hugging Face Transformers, Diffusers, JAX
- Data Handling: Pandas, NumPy, OpenCV
- LinkedIn: Varad S. Pendse
- GitHub: You're already here! 🎉 Feel free to explore my repositories.
Thank you for visiting! If you have ideas for collaboration or feedback, feel free to reach out.