Vinayak Muttathu Santhoshkumar

P: (347) 889-4996 LinkedIn | GitHub | Portfolio E: vinayakms221@gmail.com

Education:
●​ New York University, New York, NY: Masters of Science in Computer Science | GPA: 3.87 ​ ​ May 2024
●​ Vellore Institute of Technology, India: Bachelor of Technology in Computer Science Engineering | GPA: 3.58 ​ June 2022
Technical Skills:
●​ Programming Language: Python, C, C++, Java, JavaScript, Typescript, HTML/CSS
●​ Database: Oracle Database, PostgreSQL and NoSQL (MongoDB)
●​ Tools and Frameworks: SAP SuccessFactors, SAP Business Objects Web Intelligence (WebI), React, NodeJS, AngularJSExpressJS, AWS,
Git, Github, Github Actions, CircleCI
●​ Others: Software Engineering, Algorithms, Data Structures, Deep Learning, GenAI, Natural Language Processing, Image Processing
Work Experience
Business Intelligence Developer (Contract) | Valero Energy Corp - Texas, US ​ ​ Sep 2024 - April 2025
●​ Designed and optimized Oracle SQL queries, improving the report performance and reducing execution times for the learning platform
by 60%.
●​ Created and maintained integrated reports and visualizations by consolidating data from multiple sources using SAP BW/4HANA and
SAP BusinessObjects.
●​ Developed custom reports in SAP Webi, ensuring data accuracy and relevance for business needs.
●​ Enhanced database efficiency by implementing best practices in SQL query design and indexing.
●​ Streamlined reporting workflows, reducing turnaround time for data retrieval and visualization.

Machine Learning Engineer (Intern) | Tvishtryon Solutions Private Limited, India​ ​ ​ Feb 2021 – April 2021
●​ Proposed an Augmented Reality based try-on solution for customers to get a virtual in-hand experience of wearable products.
●​ Conducted performance optimization techniques to ensure real-time responsiveness and smooth user interaction.
●​ Applied PNP Ransac algorithm proficiently to extract key points to facilitate accurate transformation & mitigate occlusion challenges.
●​ Participated in regular code reviews and provided constructive feedback to team members to improve code quality and maintainability.
●​ Utilized Google MediaPipe for precise hand tracking, enhancing virtual in-hand experience.

Projects
Google Tasks Aggregator​ ​ ​ ​ ​ ​ ​ ​ ​ ​
Developed a task management tool to address deficiencies in Google Suite’s task system, aggregating assigned comments and tasks
from Google Docs, Sheets, Slides, Gmail, and Google Calendar into a single interface. Implemented features to display tasks and
comments, provide notifications for assigned and completed tasks, and convert assigned comments into Google Tasks.
Skills: React, Javascript, Python, Flask, Google Cloud, Google API
DoConnect - AI powered professional network for Doctors ​ ​ ​ ​ ​ ​
Implemented a platform for Doctors to network and collaborate along with an AI powered chatbot for efficient disease diagnosis and
patient record analysis using a Microservices based scalable and reliable architecture.
Skills: React, AWS, Sagemaker, Lex, Textract, OpenSearch, Cognito, DynamoDB, S3, and API Gateway.
Face Mask Detection using CNN and Raspberry pi
Developed a CNN model from scratch using a Kaggle dataset to detect mask usage in public spaces. Enhanced accuracy through
fine-tuning and advanced image processing techniques. Successfully deployed the solution on a Raspberry Pi for real-time monitoring.
Skills: Tensorflow, Computer Vision, Deep Learning.
Inventory Management Using Deep Learning​ ​ ​ ​ ​ ​ ​ ​
Designed and engineered a sophisticated product that provides online fashion stores an architecture to automate their inventory
management processes for categorizing apparels not only based on apparel category but also the patterns (floral, striped, checked,
plain) and brands(using a custom logo detection model).
Skills: Tensorflow, Computer Vision, Deep Learning.
Self-Steering Model for Autonomous Cars​ ​ ​ ​ ​ ​ ​
Leveraged Udacity Autonomous car data to train CNN for predicting steering angle using video frames and edge detection techniques
for increased accuracy
Skills: Tensorflow, Computer Vision, Deep Learning