Welcome !!!

Hey there! I’m a Data Scientist based in sunny San Francisco, CA.

Here is a list of my Projects, please enjoy

Bidirectional Search

Implemented Bi-Directional Search using heuristic functions, including MM0 and MM. Additionally tested their effectiveness on complex search space. text inside of a div block.

Authorship Obfuscation

Implemented an ensemble technique of SVM classifiers and the MUTANT-X obfuscator, achieving high METEOR scores and reducing obfuscation time to 1 minute. 

Proximal Training for GANs

Explored and compared proximal training with other GAN training methods, achieving high-quality images of CelebA dataset images using wasserstein GANs.

Deep Q-Learning

Implemented an autonomous navigation system using Deep Q-learning with reinforcement learning, achieving successful simulation of a car's movement in the Carla Simulator.

Camera Precision for Mono SLAM

Investigated the impact of camera resolution and field of view on MonoSLAM's precision, providing recommendations for optimal selection in various applications.

Collective Movements in Robot Swarms

Simulated the collective movement of multiple robots using MATLAB and collective localization, demonstrating expertise in multi-robot system.

6 DOF Robotics Arm

Designed, printed and controlled a 6 DOF robotic am using MATLAB and Arduino improving kinematic accuracy and proficiency in embedded systems processing.

Professional Experience

Machine Learning Engineer | PinHous | Intern

Utilize computer vision for  extracting property features from real estate images, assess data quality,  advise on project capabilities, and implement recommendation engines.  Orchestrated the setup of MySQL database on AWS RDS and devised data  pipelines for tailored recommendations.
* Developed a recommendation engine utilizing collaborative  filtering and content-based algorithms to increase customer retention by 20%.
* Engineered personalized house feeds for users based on their  preferences, search history, and engagement patterns, resulting in a 20%  increase in user interaction with recommended listings.
*Established and optimized MySQL databases for efficient storage,  retrieval, and management of extensive house listing data.
* Direct A/B testingefforts to achieve a targeted 10% increase in conversion rates, aiming for ameasurable impact on user engagement and revenue generation.

Computer Vision Engineer | ActiveIQ                                                                                                                                    

At ActiveIQ, contributed to deep learning team to  develop custom kits for vehicles to enable autonomous functionalities, with a  primary focus on commercial vehicles traveling between cities.       
* Developed lane  algorithm utilizing gradient filters, perspective transform, and custom image  segmentation techniques implemented in Pytorch to reduce offset by 30%.
* Achieved a  remarkable 96%-pixel accuracy for the assisted driving mode on Indian roads,  enhancing the vehicle’s capabilities to navigate complex and diverse road  conditions.·         
* Presented  findings to stakeholders, providing detailed reports on model accuracy, false  positive/negative rates, and computational efficiency.        
* Collaborated  with sensor fusion strategies incorporating data from cameras, LiDAR, and  radar sensors to enhance the accuracy by 20% of the perception system.

Educational Background

Master of Science – Robotics & Autonomous Systems (AI)

2021 –2023

Bachelor of Technology – Computer Science and Engineering

2015 –2019

My Medium Articles

Binary Classification for Kaggle competition: SVM, LightGBM, Decision Tree, Gradient Boosting, feature engineering, and CatBoost.
RAG and Few-Shot Prompting in Langchain : Implementation
SQL and RDBMS Crash Course— Part 1