AI and machine-learning tools I use, evaluate, or study
This is a working toolkit—not a claim that Khalil Abu Mushref created the third-party models or frameworks listed below. It covers technologies he has used, evaluated, or studied for retrieval, natural-language processing, forecasting, analytics, and healthcare-focused AI work.
Links go to each technology's primary project or documentation so readers can verify capabilities and current technical details at source.
RAG & LangChain
A retrieval-augmented generation toolkit that connects language models to selected knowledge sources. I use this pattern when exploring grounded document workflows, requirements support, and business-analysis assistants.
XGBoost & Gradient Boosting
A widely used gradient-boosting library for structured-data prediction and classification. It is part of my toolkit for forecasting experiments, KPI analysis, and research-oriented predictive modeling.
Qlib – Quantitative Modeling
Microsoft's open-source quantitative-research platform for experimenting with financial datasets, machine-learning pipelines, time-series signals, and portfolio research.
Llama 4 Maverick
A Meta language-model family I evaluate for multilingual generation, document assistance, and agent-oriented product experiments. Model specifications and availability should be checked in Meta's current documentation.
Gemini 2.5 Pro
A Google multimodal model I evaluate for text, visual, and reasoning workflows in product prototypes. Current capabilities, limits, and availability are documented by Google.
Apache Spark MLlib
Apache Spark's distributed machine-learning library for scalable data preparation, modeling, and pipeline experiments across larger datasets.
Med-PaLM 2 & Medical LLMs
A Google research model for medical question answering and clinical-language research. I study this work as context for healthcare AI; it is not presented here as a diagnostic service or a model I created.
MONAI – Medical Imaging AI
NVIDIA's open-source framework for medical-imaging research and development. It provides building blocks for imaging pipelines, while any clinical use requires independent validation and appropriate governance.
BioBERT – Biomedical NLP
A biomedical language model used in research on biomedical text mining and terminology extraction. I reference it when studying clinical-language workflows and healthcare-focused NLP.