Deeksha Makholiya Email & Phone Number

Bengaluru, Karnataka, India

- Performed exploratory data analysis (EDA) on large datasets to discern trends and patterns, enabling informed decision-making and enhancing organizational insight into key metrics.- Created interactive dashboards and reports using Tableau and visualized data using Python (Matplotlib) to present complex information in a clear and accessible manner for stakeholders.- Developed predictive data models using Python to forecast market trends with a 10% increase in accuracy compared to previous methods, providing stakeholders with more reliable insights.- Utilized Python to analyze data pulled from various sources, such as websites and Excel sheets, identifying trends and informing strategic decision-making while optimizing operational processes.

Dresden, Saxony, Germany

- Performed semantic segmentation of 50 GB of ‘.tiff’ type file satellite image data to analyze the topographic changes in the last 75 years in the land cover area of Saxony, Germany. - Provided data cleaning by enhancing the image quality by reducing noise and reducing the size of image by splitting image using Python (Numpy).- Implemented DBSCAN algorithm combined with threshold in Python for cloud masking, which helped eliminate clouds and detect high albedo land and buildings, which improved the accuracy by 40%.

Dresden, Saxony, Germany

- Developed CNN-LSTM architecture to estimate and analyze the vegetation wildfires and the area affected.- Performed comparative analysis of the developed architecture with five other ML algorithms.> Result: Implemented CNN LSTM for vegetation fire prediction in Portugal, outperforming 5 ML algorithms by 25% in accuracy. Hyperparameter optimization yielded a remarkable 30% improvement in performance metrics. Subsequently, SARIMA model excelled, surpassing CNN LSTM by 40% in future forecasting accuracy for vegetation fires.

Dresden Area, Germany

- Presented research findings to stakeholders through presentations on diverse dimensionality reduction algorithms.- Performed EDA on grayscale pixel images and developed data models for predictive analytics using Python. - Conducted a qualitative analysis between different clustering and dimensionality reduction algorithms thereby implementing autoencoders using Keras demonstrating proficiency in neural network architecture design and model training.- Developed expertise in data preprocessing techniques, including normalization and reshaping, to prepare input data for effective neural network training.

Master Of Science - Ms, Mathematics And Computer Science

Course Description: Artificial Intelligence, Machine Learning, NLP, Deep Learning, Image Processing Project : Learning OWL ontologies.

Bachelor’S Degree, Computer Science, First Class With Distinction

Activities and Societies: Team Member of AIESEC in Dehradun Course Description : Data Structures, Operating Systems, Advance Architecture.

High School, Science

Activities and Societies: House Captain -2010 Sports Vice Captain-2009 Dance Choreography