Zindi solution — multilingual maternal/reproductive health QA across 8 low-resource African languages (public LB 0.70)

SparkNLP and Healthcare SparkNLP based analysis of scientific literature on equine colic.

Clinical Reasoning Platform Base Repository

Predicting emergency room triage levels using traditional machine learning and deep learning with Turkish medical text data and structured hospital records.

CareConnect uses state-of-the-art large language models (LLMs) to provide rapid, reliable medical guidance. This project addresses increasing wait times and health misinformation, offering timely assistance and supporting informed decision-making to alleviate the burden on the healthcare system.

GRACE 'Granular Recognition of Argumentative Clinical Evidence' shared task work for IBERLEF26, with preliminary work on the CasiMedicos-Arg dataset

A role-aware RAG agent that answers the same medical question differently for patients vs. doctors -- grounded in sources, with a reproducible offline evaluation suite and CI guardrails.

Geometric Deep Learning for Clinical Informatics: Converting FHIR to Temporal-Hierarchical Graph Tensors

TrustMed-RAG: A Verification-Driven Retrieval-Augmented Generation framework. Implements dual-stage verification loops to guarantee factual accuracy and eliminate hallucinations in AI-driven medical decision support systems.

BERT-based fine-tuning for Italian Lupus clinical reports

Healthcare text analytics pipeline for clinical note processing, entity extraction, evaluation, API access, and dashboard reporting.

Herramienta de anotación clínica (SCTID/SNOMED CT) para el corpus SEMANTIAR — anotador web en TypeScript para calibración de textos clínicos en español rioplatense.

A Streamlit application designed to simplify complex medical reports into easy-to-understand language for patients. The system leverages a fine-tuned FLAN-T5 model with LoRA adapters, combined with OCR text extraction (Tesseract) and NLP preprocessing (spaCy), to deliver accurate and accessible medical explanations through a clean interface

High-performance, privacy-first medication instruction parser. Self-learning NLP engine with WebAssembly. 114K ops/sec, 8μs latency, 100% local processing. HIPAA-ready healthcare AI.

Metastases Italian Sentence Transformers Inference Classification

Llama Metastasis Instruction Tuning with LoRA

The module to extract clinical information and relations from clinical text

RAG system for drug safety Q&A over openFDA data - hybrid retrieval (BM25 + FAISS), cross-encoder reranking, grounded generation with citations and confidence-based abstention, FastAPI + Streamlit.

Reproducible end-to-end clinical NLP + MLOps pipeline for multi-label ICD-10 auto-coding on MIMIC-IV. Bronze→Silver→Gold→splits→train→evaluate, fully tested, CI-gated, DUA-compliant. TF-IDF + LR baseline shipped (test Micro F1 0.617, Macro F1 0.584); Bio_ClinicalBERT fine-tune loop validated locally; Databricks GPU run pending.

Türkçe ilaç-hastalık etkileşim değerlendirici · Turkish drug–disease interaction evaluator — LoRA fine-tuning of Turkish-Llama-8B & Mistral-7B (research/education only)