Rajshri Machine Learning Engineer
No reviews yet

Hi, I’m Rajshri, a Master’s graduate in Electrical and Computer Engineering from the University of Waterloo, specializing in Machine Learning and AI. I help students understand concepts in a simple, practical, and step-by-step way, whether you're just starting out or working on advanced projects. I can assist with machine learning and deep learning (including CNNs, Transformers, and NLP), Python, SQL, data analysis using Pandas, NumPy, and PySpark, as well as core data science concepts like EDA, feature engineering, and model evaluation. I also provide guidance on academic assignments and projects. My experience includes working with large-scale industrial datasets (60M+ records), building end-to-end ML pipelines, and deploying solutions on cloud platforms such as AWS and Azure. I focus on explaining the “why” behind concepts, not just the theory, and my teaching style is beginner-friendly, using real-world examples and coding walkthroughs to ensure structured learning from basics to advanced. Additionally, I can help with school-level math and physics for grades 8–12, as well as IELTS English preparation. Whether you need doubt-solving, assignment guidance, or want to build strong fundamentals, I’d be happy to help!

Subjects

  • IELTS

  • Maths (12th)

  • Physics (10+2)

  • Machine learning projects Beginner-Expert

  • Machine and Deep Learning project Beginner-Expert

Experience

  • Graduate Research Assistant (May, 2024–Dec, 2025) at Univeristy of Waterloo/ Waterloo
    I have worked as a Machine Learning Researcher at the University of Waterloo, where I contributed to two significant research projects. In my first role (March 2024 – December 2025), I focused on my thesis titled “Study of Deep Learning Architecture for Bearing Fault Diagnosis using STFT Spectrograms” under the supervision of Dr. Kshirasagar Naik and Dr. Mahesh Pandey. I engineered an end-to-end deep learning pipeline that integrated signal preprocessing, model training, hyperparameter tuning, and rigorous evaluation across ten imbalanced classes. I developed custom 1D CNN architectures for signal modeling and applied transfer learning by fine-tuning on 2D and 3D spectrogram representations, while cross-validating robustness through different datasets, ultimately reducing training time by 25%. Through this work, I achieved a 0.99 macro F1-score and ROC-AUC across fault classes by leveraging transfer learning and fine-tuning of pretrained models.

    In my second role (May 2024 – August 2024), I worked on Behaviour Monitoring of Induction Motors in collaboration with BRUCE POWER, again under the guidance of Dr. Kshirasagar Naik and Dr. Mahesh Pandey. I developed an unsupervised anomaly detection pipeline on over 62 million industrial records, which helped reduce unplanned downtime risk and maintenance costs under variable load cases. My work reduced manual diagnostic workload by 40% using Isolation Forest and DBSCAN, paired with interpretable visualizations through PCA and t-SNE. I also built a scalable diagnostics workflow on Azure Databricks, enabling rapid retraining and seamless transfer across nuclear facility test environments.

Education

  • Masters in Applied Sciences (Sep, 2023–Dec, 2025) from University of Waterloo, Waterloo, Canada–scored 3.7

Fee details

    CA$30–45/hour (US$21.74–32.61/hour)

Reviews

No reviews yet. Be the first one to review this tutor.

© 2016 - 2026 TeacherOn. All rights reserved