• Pratik Ambadskar
• Nirbhay Singh

The rapid advancement of financial analytics has enabled stock market prediction using machine learning and sentiment analysis from social media and financial news. However, existing approaches often neglect critical issues of data privacy and secure deployment. This paper proposes a Privacy-Preserving Serverless Architecture for stock market prediction, integrating deep learning models with sentiment analysis while ensuring privacy of sensitive financial and user data. Differential privacy mechanisms are applied to protect user identifiers and sentiment features, while federated learning concepts are incorporated to minimize raw data sharing. The system leverages serverless platforms such as AWS Lambda and Google Cloud Functions to achieve scalability, cost efficiency, and fault tolerance. Experimental results show that hybrid deep learning models combining LSTM with sentiment features outperform classical baselines, reducing error rates by 10–15%. Furthermore, privacy-preserving techniques successfully protect sensitive inputs while maintaining competitive model accuracy. This study highlights the potential of secure, scalable, and privacy-aware stock prediction systems for financial technology applications

Privacy-preserving analytics; Differential Privacy; Federated Learning; Serverless Computing; Stock Market Prediction; LSTM; Sentiment Analysis

"PRIVACY-PRESERVING SERVERLESS ARCHITECTURE FOR STOCK MARKET PREDICTION ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 9, page no.d499-d502, September-2025, Available :http://www.jetir.org/papers/JETIR2509367.pdf

"PRIVACY-PRESERVING SERVERLESS ARCHITECTURE FOR STOCK MARKET PREDICTION ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 9, page no. ppd499-d502, September-2025, Available at : http://www.jetir.org/papers/JETIR2509367.pdf