Manish Kakarla

I am a highly ambitious Masters student in Computer Science seeking a software developer role where I can leverage my skills and experience in data structures, algorithms, problem-solving, and machine learning to make a meaningful impact as part of a team of experienced technologists, quantitative researchers, and traders. I am eager to learn new technologies and create innovative solutions that directly impact the business.

Distinguished Pioneer Scholar, CSR Scholar's Program, California State University, East Bay (2021-2022)

• Research focused on the development of machine learning accelerator techniques to improve the performance of deep learning models under the guidance of Dr. James Tandon. My name can be found on the list of CSR Scholars for the 2021-2022 academic year at the following link: https://www.csueastbay.edu/csr/past-scholars/csr-scholars-2021-2022.html

Employee of the Month, Recreation and Wellness Center

Summer 2021

• Manish joined the RAW team over the summer when we reopened after being closed for over a year. He provided enhanced levels of support, often going above and beyond to help in any way shape or form he could. He always steps in and up and asks what more can be done. He never leaves a task undone or completed half way. His work ethic is unparalleled. Thanks Manish for all of your contributions to the RAW and your commitment to our Principles of Leaders, holding yourself accountable to your role and communicating needs and creating a solid experience for our members.

• AWS machine learning foundation Nano degree by Udacity
• Deep Neural Networks by one forth labs
• Deep learning specialization by DeepLearning.ai
• Advanced TensorFlow Techniques by deeplearning.ai
• Data Scientist with python by Data Camp
• Code Path Advance Software Engineering

Travelers salesman Optimization Using Algorithms
Python, Functions, objects, NumPy, probability

• To show the shortest traveling path for 27 cities in California using different machine learning algorithms
• Used genetic algorithm to randomly populate the path at first and improve the distance through several generations
• Used Ant colony Optimization to get the path by choosing the probability of next city using the pheromone intensity and ants

Agriculture Vision Application

Python, Pytorch, Pytorch lite, Android Studio

• Detecting the Crop Anomalies From the drone images using Computer vision.
• Worked Downloading the Huge Data of Drone images 15GB with 11 classes from AWS
• Trained the data with Self constructed graph model in PyTorch.
• Build an android application using the PyTorch mobile so that it segments the given drone images to detect the Anomalie

Mask detection
August 2021 – December 2021
TensorFlow, Python, Tensor Board, Object detection, Pandas

• Detecting whether a person is wearing a mask or not Using EfficientDetD0 and object Detection API of TensorFlow
• Mounted the google drive to load the data, preprocessing the dataset like converting the xml to csv, validating annotations.
• Installing the object detection API in google Colab, generating the TFRecord, Training the model.
• Generated the results and installed tensor board to view the results of the stored records

Improving mobilenetv1 using Improved Pointwise convolutions

• The thesis explains the optimization of pointwise convolutions to reduce the number of operations and parameters in deep learning models
• Focuses on the small model, mobilenetv1, and its 94% of total parameters due to pointwise convolutions
• Utilizes a decomposition technique to optimize pointwise convolutions and reduce the number of parameters
• Results in a reduction in the number of parameters tested on classification of the CIFAR-10 dataset and object detection on Pascal VOC 2007 and 2012 (20 classes) with a reasonable loss of metrics.