Puneeth Raviprakash

Automatically identify and flag duplicate or highly similar Power BI slides

Developed an AI-driven slide deduplication system to identify and remove redundant PowerPoint slides using visual, structural, and textual similarity analysis. Implemented automated MLOps workflows for model versioning, containerized deployment (Docker, Azure ML), and CI/CD integration (Azure DevOps), with continuous monitoring and logging via MLflow and Application Insights to track accuracy and drift over time. Leveraged CLIP (Contrastive Language–Image Pretraining), SSIM, OCR (pytesseract), and all-mpnet-base v2 embeddings from Hugging Face to perform multi-modal similarity detection across large-scale slide repositories.

Assess clients' progress across Cloud, Blockchain, Cybersecurity, AI/ML, Data Maturity, and Digital Transformation stages.

Leverage Google Cloud and Generative AI tools to optimize blockchain implementations, strengthen cybersecurity frameworks, enhance AI/ML capabilities, and conduct data maturity assessments. This approach aims to refine cloud strategies, automate governance, support skills development, and streamline pilot workload selection—ultimately accelerating digital transformation across foundational, operational, and advanced maturity stages.

• ➢ Performed Data Collection, Data Preparation, Feature Engineering, Hypothesis Testing, Dimensionality Reduction and Data Mining to develop mathematical and statistical models.

• ➢ Data Pre-processing (Missing Value Treatment, Outlier Detection, Data Exclusion, Feature Engineering).

• ➢ Analysed structured data to develop and apply new and existing models and publish well documented, reproducible work.

• ➢ Transformed visualizations into meaningful insights for reporting.

• ➢ Developed solutions using statistical techniques, machine learning (Linear Regression, DIM RED: PCA, Regularization, Cross Validation), Artificial Intelligence, Sequential models (LSTM, Bidirectional LSTM, GRU), Naïve Bayes, Logistic Regression

• ➢ Performed Data Collection, Data Preparation, Feature Engineering, Hypothesis Testing, Dimensionality Reduction and Data Mining to develop mathematical and statistical models.

• ➢ Data Pre-processing (Missing Value Treatment, Outlier Detection, Data Exclusion, Feature Engineering).

• ➢ Analysed structured data to develop and apply new and existing models and publish well documented, reproducible work.

• ➢ Transformed visualizations into meaningful insights for reporting.

• ➢ Developed solutions using Ensemble techniques – CART, SVM and Random Forest