Omkar Garad

Experience

• 

  IBM

  May 2022 - Aug 2022

  Machine Learning Intern

  At IBM Research, I was part of a team aimed at predicting potential revenue-generating business opportunities with a focus on estimating their likelihood. As the sole ML intern in my team, I was in charge of the entire Machine Learning pipeline. I trained an XGBoost classification model using 10-fold cross-validation across 100 distinct splits, ultimately achieving an impressive average F1 score of 86%.To enhance the model's performance, I augmented our dataset by generating synthetic business opportunity data using Mockaroo. Furthermore, I conducted comprehensive feature extraction and engineering to identify crucial factors contributing to revenue-generating business opportunities and integrated SHAP (SHapley Additive exPlanations) to gain insights into the model's decision-making process. Show less

• 

  Ghent University

  Aug 2022 - Dec 2022

  NLP Research Intern

  ∘ Designed and implemented a multilingual complex Named Entity Recognition (NER) system capable of identifying intricate and short-context named entities across four different languages∘ Used XLM-RoBERTa for initial entity span detection and subsequently utilized domain-specific XGBoost models to categorize the entities with fine-grained tags based on their domains

• 

  Honeywell

  May 2024 - now

  AI/ML Engineering Intern