Skip to content
View billodalroy's full-sized avatar
  • Buffalo, NY
  • 03:30 (UTC -04:00)

Block or report billodalroy

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
billodalroy/README.md

Billodal Roy πŸ‘¨β€πŸ’»

About Me πŸ™‹β€β™‚οΈ

High performing data science professional πŸ“Š with over 3 years of experience in the analytics and data science domains. Proven track record in statistical modeling, machine learning and developing production-ready 🏭 ML models. Specialized in communicating complex technical concepts and translating analytics into business value πŸ’΅ across diverse stakeholders.

Experience πŸ’Ό

Data Scienctist @ Lowes Companies Inc 🏠

Data Science Intern @ Lowes Companies Inc 🏠

Data Scientist - III @ Protiviti India Member Firm 🏒

Data Science Researcher @ Stones2Milestones Edu Services 🏫

Certifications πŸ…

  • AZ-900: Azure Cloud Fundamentals
  • Google Data Analytics Specialization

Projects πŸ’»

Stack: Python, PyTorch, Gymnasium

  • Developed deep neural network agent using PyTorch to play classic Snake game, leveraging algorithms including DQN, Double DQN, and A2C to optimize policy. Achieved score 10x higher than baseline agent. πŸπŸ€–πŸ†

  • Engineered features and reward functions enabling agent to master Snake game environment. Compared performance of DQN, DDQN, and A2C algorithms in optimizing gameplay policy, identifying most effective techniques.

Flight Delay Prediction

Stack: Python, SQL, REST API

  • Built a two-stage machine learning model using Random Forest algorithms for classification and regression to predict flight arrival delays, achieving an F1 score of 0.78 and RMSE of 11.28 minutes. βœˆοΈβŒšπŸ“ˆ

  • Preprocessed imbalanced flight data using SMOTE oversampling before training Random Forest models to optimize performance and reduce prediction error compared to baseline models. βœˆοΈπŸ“‰πŸ“ˆ

Pinned Loading

  1. fselect fselect Public

    A fast and scalable implementation of Feature Selection for Clustering

    Python 2

  2. Snake-Game-Deep-Reinforcement-Learning Snake-Game-Deep-Reinforcement-Learning Public

    Teaching AI to learn the Snake Arcade game

    Jupyter Notebook

  3. ub-iss-chatbot ub-iss-chatbot Public

    A Mistral 7B LLM powered chatbot that answers queries related to international students at University at Buffalo using context documents

    Python