Jatin A. Salve

📍 Gainesville, FL, USA

📧 jatin.salve@ufl.edu

📞 +1 (352)-757-9671

I’m Jatin Salve, a Computer Science graduate student at the University of Florida who builds trustworthy AI systems for healthcare, accessibility, and large-scale decision making. I enjoy bringing research ideas into production environments where safety, latency, and interpretability matter.

Research & Industry Highlights

Designed high-throughput ML services at ICICI Bank, delivering real-time fraud and personalization signals for 10M+ customers.
Led evaluation and verification subsystems for medical LLMs, translating RDF query results into human-auditable evidence.
Published multimodal clinical summarization work (ACL 2024) that outperforms GPT-3.5 and GPT-4V on factual recall.

Technical Stack

AI/ML: Python, PyTorch, TensorFlow, JAX, LangChain, FAISS
Knowledge Systems: RDFLib, SPARQL, graph databases, retrieval-augmented generation
Distributed & Product: Docker, Kubernetes, AWS, Linux, React.js, Node.js
Experimentation: Weights & Biases, Apptainer, HiPerGator HPC

Community & Mentorship

Google Summer of Code alumnus and long-time JoomlaCMS contributor, with upstream modules adopted by thousands of sites.
Mentor for underrepresented students learning machine learning and open-source development.
Advocate for applied AI in healthcare, presenting workshops and lightning talks at academic and community meetups.

Let’s Collaborate

I’m always excited to explore research collaborations, product partnerships, or speaking opportunities around multimodal AI, clinical NLP, and responsible deployment. Dive into my publications, explore ongoing projects, or grab my CV. Feel free to reach out—let’s build something meaningful together.

news

Aug 15, 2024	🎉 ACL 2024 Publication: Our paper “From Sights to Insights: Towards Summarization of Multimodal Clinical Documents” has been accepted at the 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024)! This work presents EDI-Summ, a novel multimodal summarization architecture that outperforms GPT-3.5 and GPT-4V on clinical summarization tasks.
May 01, 2022	MITACS Globalink Research Internship at Polytechnique de Montréal
Jul 01, 2021	🚀 Google Summer of Code 2021: Excited to announce my selection for Google Summer of Code with JoomlaCMS! I’ll be developing a customizable guided tour system to improve user onboarding experience. This highly competitive program (0.2% acceptance rate) will allow me to contribute to open source while working with an amazing global community!

selected publications

ACL
From Sights to Insights: Towards Summarization of Multimodal Clinical Documents

Jatin Avinash Salve and others

Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL)This work represents a significant advancement in multimodal clinical NLP, with applications in healthcare documentation and medical decision support systems. , Aug 2024

Abs Bib

We present EDI-Summ, a novel multimodal summarization architecture that combines contextual and cross-attention mechanisms for clinical document summarization. Our approach achieves state-of-the-art performance with 0.81 factual recall and 3.51 clinical score, outperforming GPT-3.5 and GPT-4V on clinical summarization tasks. The system integrates visual and textual information from clinical documents to generate comprehensive summaries that maintain both factual accuracy and clinical relevance.
@article{salve2024multimodal, title = {From Sights to Insights: Towards Summarization of Multimodal Clinical Documents}, author = {Salve, Jatin Avinash and others}, journal = {Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL)}, year = {2024}, month = aug, publisher = {Association for Computational Linguistics,}, }