Machine learning researcher at Medtronic Surgical R&D | PhD candidate at University of New Haven
My research focuses on clinically robust vision-language models for diagnostic applications, specifically addressing phrasing-sensitive failures and misleading explanation effects in medical AI systems.
- PhD dissertation on clinically robust vision-language models for diagnostic reasoning
- Investigating coupled failure modes in medical VLMs using MIMIC-CXR
- Aegis: an AI-powered personal operations system integrating fitness and productivity data
- Mechanistic interpretability and causal mediation analysis in VLMs
- Parameter-efficient fine-tuning (LoRA) for medical imaging models
- Cross-attention mechanisms in multimodal architectures
- Medical AI and healthcare ML applications
- Vision-language models for clinical diagnostics
- AI safety frameworks for medical settings
- Medical AI and vision-language models
- Chest radiograph analysis
- ML in surgical applications
- Running and endurance training
Training for a sub-20 minute 5K while chasing a 1,000-mile annual running goal
- Email: contact(at)bineshkumar(dot)me
He/Him