SUMAN BEZBARUAH

Guwahati, India, +918638781993, sumanbezbaruah1234@gmail.com, http://www.linkedin.com/in/suman0007/, https://github.com/sumam121

• Microsoft Azure Fundamentals: Describe cloud concepts, https://drive.google.com/file/d/1b9oGvZye65_TN9pqmfWRo3vQllKFjLxY/view?usp=sharing
• Microsoft Azure Fundamentals: Describe Azure architecture and services., https://drive.google.com/file/d/1b9oGvZye65_TN9pqmfWRo3vQllKFjLxY/view?usp=sharing
• ML internship, https://drive.google.com/file/d/1NESwJEPQlTAuB9N_vpaHGM0MHZ6k9LhM/view?usp=sharing
• Core and Advance Java, https://drive.google.com/file/d/16NbWx4xL6BXSk8A0E3hZBl5l4Ma_YJg-/view?usp=sharing

• English
• Hindi
• Assamese

• Neural Network for Handwritten Digits Classification

 02/2024, 02/2024

• Pre-processed and normalised the MNIST dataset for training and testing our MLP model. Classified handwritten digits with 91.7% accuracy using a multi-layer perceptron neural network
• Calculated loss using Sparse categorical Cross-entropy and used Sigmoid activation function. Fine-tuned hyperparameters through iterative testing to improve the model's predictive capabilities.
• Face Recognition

01/2024, 01/2024

• Data preparation: I gathered a diverse dataset of faces from various individuals.
• Detect and Recognize: I used the Haar Cascade Classifier to detect faces and the Local Binary Pattern Histogram (LBPH) algorithm to recognize faces.
• Mapping Functions: I designed custom functions to map person names to numeric labels and vice versa, enabling seamless processing of training data.
• Training and Evaluation: The model was trained on the prepared dataset, achieving impressive accuracy of 82% in recognizing known faces.

 Stock Price Prediction using LSTM 

09/2023, 10/2023

• Fetched historical stock data from Yahoo Finance of tesla Stock.
• Building and training LSTM(Long Short Term Memory) Deep Learning Model models using Keras.
• Predicting future stock prices and evaluating model performance. Visualizing actual and predicted prices using matplotlib.