Sumit Kothari Email & Phone Number

Los Angeles, CA, US

Los Angeles, California, United States

• Developed comprehensive full-stack portals for USC - Alzheimer’s Therapeutic Research Institute (ATRI), utilizing ReactJS for the frontend and Django for the backend with PostgreSQL, ensuring responsive design and.
• Implemented Test Driven Development (TDD) and Agile methodologies, employing Cypress for frontend testing, Selenium for backend testing, and Jira for task management to achieve high reliability and performance.
• Created a CI/CD pipeline using GitHub Actions to automate frontend, backend, and functional tests before deployment, enhancing the overall development workflow and deployment efficiency.

• Developed a LLM Text2SQL model to generate and execute SQL queries on custom databases using the GPT-3.5-turbo model and Azure OpenAI embeddings, incorporating client-specific information and business rules for.
• Led backend development with tools such as FastAPI, Pydantic, and SQLAlchemy, and managed the team codebase using Azure DevOps for efficient collaboration and version control
• Engineered a CI/CD pipeline for seamless deployment of the bot on Azure Cloud, integrating it with MS Teams, implementing Single Sign-On, and enhancing the chat UI with TypeScript.

Mumbai, Maharashtra, India

• Created a federated learning model using PySyft and PyTorch to improve the privacy and security of the model. This model was able to detect anomalies in Bitcoin transactions by leveraging the power of federated learning.
• Employed Truffle to develop a custom dataset with transactions through Ethereum and track all attackers. This involved working with blockchain technology to collect and store transaction data for analysis.
• Utilized Tensorflow to train several Deep learning models such as SVM, Autoencoder, variational autoencoder, and deep belief network for anomaly detection in Bitcoin transactions. These models were tested and optimized.

• Devised a system that leveraged computer vision techniques to detect license plates and capture the relevant information.
• Worked with OpenCV to implement advanced computer vision techniques such as reducing noise in images, edge detection, and masking to improve the accuracy of license plate detection.
• Utilized state-of-the-art machine learning techniques such as Yolo and DeepSORT to detect and track vehicles in real-time. I also utilized EasyOCR to read the numbers on the license plate. This helped automate the.

• Collected and analyzed Reddit posts from the 'depression' and 'SuicideWatch' subreddits using the Pushshift.io API Wrapper (PSAW). This involved scraping around 30,000 posts, cleansing and classifying them, and.
• Compared and trained multiple machine learning models including SVM, Logistic Regression, Naïve Bayes, and ANN to predict whether a person is experiencing depression. This involved selecting and tuning the best model.

Master Of Science - Ms, Computer Science

Bachelor Of Engineering - Be, Computer Engineering, 9.55