Blogs & Papers

Introducing BODHI — an open clinical knowledge graph for grounding healthcare AI in verified, India-specific clinical data.

Releasing the gold standard for medical speech recognition in Indian healthcare.

Bridging India's multilingual healthcare gap with multilingual embeddings and benchmarks.

An open-source framework for evaluating medical AI systems.

New evaluation metrics for speech recognition in healthcare contexts.

Purpose-built vision LLM for parsing and extracting data from medical documents.

Purpose-built LLM for English ASR in Indian healthcare settings.

EkaCare releases four evaluation datasets for medical AI systems.

Comparative analysis of specialized small models versus general-purpose AI for healthcare tasks.

Challenges and learnings from extracting structured data from medical lab reports.

Ontology applications in medical contexts for knowledge representation.

EkaCare's annual data-driven health report for India.

Time series analysis of NYC 311 noise complaint data to measure urban activity recovery post-COVID.

NLP pipeline to classify and extract signal from Glassdoor employee reviews — identifying sentiment, topics, and leading indicators of company health.

Graph ML approach to mapping competitive dynamics among private companies using web and textual signals.

Real-time tracking and forecasting of US electricity demand as an alternate economic indicator.

Doctoral thesis on melody-centric computational models for Indian classical music — tonic identification, raga recognition, melodic similarity, and pattern discovery.

Master's thesis on automated tonic identification for Indian classical music using signal processing and machine learning.