Prachi Jirimali

About Prachi Jirimali

Machine Learning Engineer with 4+ years of hands-on experience in developing and deploying AI/ML solutions across Natural Language Processing (NLP), Computer Vision, and Predictive Analytics in BFSI and Tech sectors.• Proficient in Python and deep learning frameworks such as TensorFlow, PyTorch, and Keras with expertise in designing and fine-tuning CNNs, LSTM networks, Transformer architectures (BERT, GPT-2, RoBERTa), and GANs.• Experienced across the ML lifecycle: data wrangling and feature engineering (NumPy, Pandas), model development and evaluation (Scikit-learn, TensorFlow, XGBoost, SHAP), and deployment (Flask, Docker, AWS, Azure).• Well-versed in NLP tasks including sentiment analysis, NER, text classification, intent recognition, and topic modeling using spaCy, NLTK, and Hugging Face Transformers.

Experience

1. Machine Learning Engineer

   PNC

   Jan 2024 — Present

   Developed and fine-tuned Transformer-based NLP models (BERT, GPT-2) for customer sentiment analysis, intent classification, and FAQ retrieval, achieving over 92% accuracy and improving automation in customer interactions.• Engineered a Generative Adversarial Network (GAN) to synthesize diverse customer profiles for training fraud detection models, enhancing rare-event recall by 18% and improving model generalization.• Built a Generative AI module integrating NER, sentiment tagging, and topic modeling using TensorFlow, spaCy, NLTK, and RoBERTa; achieved >90% accuracy on validation datasets across modules.• Led the development of an AI-powered customer support chatbot using NLTK and fine-tuned GPT-2, automating over monthly queries and reducing average response time by 30%.• Designed a Random Forest-based churn prediction system with SHAP explainability; F1 score of 0.92 led to a data-driven retention strategy, decreasing customer churn by 11%

Education

• New Jersey Institute of Technology

  Master's degree

• Terna Engineering College

  Bachelor of Engineering - BE

  2016 — 2020