Created a speech clutter disorder classification system for 3,000 records using acoustic feature vectors obtained from audio-

Conducted digital signal processing and trained a deep neural network (DNN) model with K-fold validation using Python, Librosa, TensorFlow, Torch, and Keras, achieving 85% accuracy on test records.

Developed and implemented a Q-Learning algorithm to solve reinforcement learning problems in the Open AI Gym environment.

Trained an agent over 100,000 episodes using an epsilon-greedy policy and updated the Q-table using the Bellman equation.

Created a Python script using data analytics to rank Tennis ATP players depending on player stats. The client rated the script "extremely useful."

Created a machine learning algorithm to forecast supply needs in a busy restaurant based on past receipts reducing inventory.

Designed and developed a system using Jackal Robot for object detection and surrounding rendering to accurately reflect the

Setting up the 3-way cross-platform interaction between Jackal (ROS) to Python and Python along with the PIFUHD model.

Designed an architecture pattern of python framework which can be used as plug and play for different data analyst engineer.

Designed, implemented, and deployed the complete workflow of geometric shape generation along with the mentorship of two interns

Worked on optimization of YOLOv3, Zero-Short-Learning, and GANs algorithms and reduced their training time by 50%.

Created Python scripts to perform ETL processes using Glue, Lambda, and Sage Maker.

Built Excel and Tableau dashboards, tools, and ad-hoc reports to improve reporting, analytics, and visualization needs while