Developed and evaluated a Transformer-based NLP model for movie review sentiment classification. The study applies tokenization, attention mechanisms, and model fine-tuning to classify reviews as positive or negative, benchmarking a full encoder-only Transformer pipeline against statistical baselines.

AWS AI & ML Scholars · AWS Future AI Programmer · Technical Report

Engineered a credit scoring model to predict loan default risk from financial and repayment history. The research compares classical classifiers and ensemble methods, using feature engineering and threshold optimization to support risk assessment protocols.

CodeAlpha · Technical Report

Architected a Convolutional Neural Network for recognizing handwritten digits and characters from images. The work covers image preprocessing, data augmentation, and CNN architecture design, with an extension path to full word recognition using sequence models such as CRNNs.

CodeAlpha · Technical Report

Investigated the key drivers of air ticket booking behaviour through an end-to-end machine learning pipeline. The research also develops a scalable demand forecasting framework supporting capacity planning for premium lounge departures at a major hub airport.

British Airways · Virtual Experience Program · Research Brief

A systematic literature review of evaluation frameworks, benchmarks, and practical assessment methodologies for large language models. The review synthesizes prior taxonomies to map reliable, reproducible evaluation practices for production and agentic LLM systems.

Algoverse AI Research · Algoverse Research Group

A technical investigation of design patterns for agentic AI systems, covering tool-using agents, subagents, and Multi-context orchestration. The research analyzes how structured protocols such as the Model Context Protocol enable secure, composable interactions between models and external systems.

NexusTrade · Anthropic · Technical Exploration