Deep Learning Computer Vision

Vastav Bharambe

470-***-**** # **********@******.*** ï linkedin.com/in/vastav06 § github.com/vastav06 Education

Georgia Institute of Technology Atlanta, GA

Master of Science in Artificial Intelligence and Robotics GPA: 3.9/4.0 Aug 2022 - May 2024 Graduate Teaching Assistant: Big Data Systems, Guide: Ling Liu Symbiosis Skills and Professional University Pune, India Bachelor of Technology in Mechatronics GPA: 8.99/10 Aug 2018 - Jul 2022 Technical Skills

Languages: Bash, Python, C, C++, MATLAB, R, Javascript, SQL, HTML, CSS Tools: React, Tensorflow, Kubernetes, Tailwind, Bootstrap, PySpark, Numpy, Pandas, Scikit-learn, Optimization, NLTK Technologies/Frameworks: OpenCV, Docker, Git, Linear Regression, Time-Series, Power BI, Pytorch, React.js, AWS Work Experience

Learnmutiny May 2023 - Aug 2023

Deep Learning Intern Atlanta, GA

• Designed a 6-page CRM platform using Tailwind and JavaScript, resulting in a 40% boost in recruiter efficiency

• Achieved 20% faster candidate sourcing by developing a Chrome extension and storing the records using AWS-RDS

• Integrated LDA and cosine similarity using an LLM model for API recommendations and harnessed Gensim for user-to-user topic modeling to achieve 95% accuracy Georgia Tech and Temple Allen Industries Jan 2023 - May 2023 Computer Vision Research Assistant, Guide: Dr. Jacob Abernethy Atlanta, GA

• Collected live 3D Point Clouds for airplane sanding operation using ROS bags to perform ICP and CPD analysis

• Minimized the mapping error from 4 cameras using YOLOv7 for end-effector detection in an occluded setting

• Enhanced pose robustness by combining ICP with 6D DensFusion, PointNet, and CenterSnap to achieve 90% accuracy Automation Edge Aug 2021 - Feb 2022

Machine Learning Engineer Intern Pune, India

• Engineered RPA workflows and web scraping tools in JavaScript to optimize and analyze NSE2 financial market data

• Optimized a Large Language model by 15%, enabling the AI CogniBot to accurately interpret semantic nuances Omdena Oct 2021 - Jan 2022

Machine Learning Engineer Intern New York, NY

• Trained a YOLOv5 model within a Computer Vision pipeline to highlight features like the fairness of 6 ethnic races

• Executed image synthesis techniques to identify key features and handled racial bias to achieve an accuracy of 96%

• Led a team of 20 to deploy the model on a live news feed using dockerized containers to a European startup Eklavya Infosys, in collaboration with ThoughtWorks Jun 2020 - Jul 2020 Data Science Intern Pune, India

• Developed time series forecasting algorithms for inventory management on a 10-year demand-driven Walmart dataset

• Constructed an end-to-end ML pipeline to examine ARIMAX, XGBoost, and Random Forest, achieving 97% accuracy Projects

VGG19 based Skin Lesion Classification with Deep Learning and Image Processing Tensorflow, Pytorch

• Improvised the generalization ability of classification model using data augmentation techniques over 7 Lesion Classes

• Proliferated the scalability of the model using VGG19 to achieve 99.04% accuracy and published a paper in IEEE Integrated Multi-Sensor Fusion and Adaptive Navigation for Optimal Maze Traversal Pytorch, CNN, ROS

• Devised a sequential neural network using IMU sensor data achieving 95% accuracy in detecting traffic signs

• Implemented a state machine incorporating angular-linear velocity controllers to navigate the maze in 150 seconds Spectro-Temporal Analysis and Optimization of Bilingual Text-to-Speech Synthesis DSP, Regex, Python

• Performed a closed loop bilingual English-to-Marathi TTS using NLP and Regex, involving 20 human evaluators

• Analyzed the spectro-temporal representation of bilingual speech and reduced the speech time from 9.69 to 2.86 seconds Robust 3D Point Cloud Classification with Transformers and Enhanced Adversarial Defense Deep Learning

• Assessed the efficiency and scalability impact by comparing 3D-CTN and SimpleView using ModelNet40

• Applied uniform sampling and noise to 3D-CTN, increasing accuracy by 15% and enhancing reliability for 3D points Publications

Border Security System for Intrusion Detection using Robotics System, IJSRET 2021 Design and Development of Aerial and Under-Water Drone for Security and Surveillance, IJISET 2021

Contact this candidate