PIYUSH AGARWAL

An Accomplished, determined and well-rounded individual with more than 2 years of experience as Data Scientist within Finance and Automotive industry. Focused and keen concentrated professional in Data Science, Machine Learning, Deep Learning, Natural Language Processing, Computer Vision and Generative AI. The contribution has been pivotal in interpreting complex datasets, developing and implementing data science models and algorithms, applying machine learning and deep learning techniques, utilizing natural language processing methods, and collaborating with cross-functional teams to translate business requirements into analytical solutions that drive significant impact.

• Machine Learning - Coursera : Relevant coursework - Supervised Learning Algorithms, Unsupervised Learning Algorithms, Exploratory Data Analysis Best Practices
• Neural Networks and Deep Learning - Coursera : Relevant coursework - Mathematics behind Neural Networks, Working of Neural Nets, Forward Propagation, Backward propagation
• Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - Coursera : Relevant coursework - Activation Functions, Loss Functions, Optimizers, Hyperparameter Tuning, Regularization for improving performance of Neural networks
• Sequence Models - Coursera : Relevant coursework - Introductions to Recurrent neural networks and journey to attention mechanism
• Data Science & Machine Learning Training InternShala : Relevant coursework - Practical Learning of Machine Learning and Data Exploration

Guard-Railing LLM Applications

• Implemented safety measures like Misinformation, Hallucination, generating correct, truthful, and consistent outputs.

• Ensured structured outputs using Pydantic data models and Rail Specs of GuardRails AI.

Sentiment Analysis

• Implemented three common word embeddings in NLP: Word2Vec, Tf-Idf and BERT.

• Trained and evaluated fine tuned BERT architecture with custom output layer.