Yadu Krishna

Postgraduate student specializing in Artificial Intelligence with expertise in Machine Learning. Demonstrated proficiency in machine learning algorithms and data preprocessing through hands-on projects. Skilled in Python and relevant libraries for developing AI models. Engaged in academic projects covering machine learning, computer vision, and NLP, enhancing analytical capabilities.

• Earned the Deep Learning Specialization Certificate from DeepLearning.AI on Coursera, comprising modules on Neural Networks, Hyperparameter Tuning, CNNs, Sequence Models, and ML Project Structuring.
• Completed a 12-month advanced AI/ML course by the Additional Skill Acquisition Programme (ASAP), an initiative by the Government of India, gaining comprehensive knowledge in Artificial Intelligence and Machine Learning
• IBM AI Engineering Professional Certificate focused on developing, fine-tuning, and deploying deep learning and LLM-based models, with hands-on experience in Generative AI, NLP, LangChain, RAG, TensorFlow, PyTorch, and Keras.
• Codecademy's Learn the Basics of Machine Learning course provided foundational skills in supervised and unsupervised learning, model assessment, and applied machine learning situations.

• Advanced Vehicle Price Prediction Using Machine Learning :
  Built an end-to-end regression system to predict vehicle prices using a large automotive dataset. Performed data preprocessing, missing value imputation, outlier handling, feature engineering, categorical encoding, and PCA-based dimensionality reduction. Trained and evaluated Ridge, Random Forest, Gradient Boosting, and Voting Regressor models with k-fold cross-validation, achieving best performance with Random Forest (R² ≈ 0.98). Applied SHAP and partial dependence analysis for model interpretability. Implemented using Python, Pandas, NumPy, Scikit-learn, and Matplotlib.

• Stock Price Prediction with Time Series Analysis :
  Developed a stock price prediction application using time series forecasting techniques, including ARIMA, Prophet, and LSTM models. Implemented real-time stock monitoring, portfolio analysis, and backtesting using yfinance data, with Plotly for interactive visualizations.

• Automated Credit Card Approval Predictor :
  Built a machine learning–based system to automate credit card approval decisions, improving processing efficiency and reducing manual errors. Analyzed key applicant features such as income, credit history, and loan balance to support accurate approval predictions.

• Optimized Retinal Disease Detection Using Deep Reinforcement Learning :
  Developed a deep learning–based medical image classification system for multi-class retinal disease detection using CNNs and transfer learning. Designed and compared a baseline CNN, a VGG19 model, and an optimized CNN enhanced with a Deep Q-Learning (DQN) agent for dynamic hyperparameter tuning. Implemented reinforcement learning with experience replay and epsilon-greedy policy to optimize filters, kernel size, pooling, dense units, batch size, and epochs. Achieved ~89–90% validation accuracy, outperforming standard CNN and VGG19 models. Implemented using Python, TensorFlow/Keras, NumPy, and Matplotlib.

• ResearchReverse – AI Research Paper Summarization & RAG System :
  Developed a local AI-powered research assistant to ingest academic PDFs, perform semantic chunking and embedding, and generate structured multi-stage summaries using local LLMs via Ollama. Implemented vector-based semantic search and Retrieval-Augmented Generation (RAG) for paper-grounded Q&A, with an interactive Streamlit interface for exploration, chat, and project management.