Hey, I'm Sudarshan Anand
As a Georgia Tech graduate with dual Master's degrees, my expertise bridges the gap between mathematical rigor and scalable machine learning solutions. Previously, I worked as a Graduate Student Researcher at the Georgia Tech Bio-MIBLab developing multimodal models to track neurodegenerative disorders, and at AdityaLab where I developed open-source library for researchers to smoothly integrate time-series foundational models in their workflows.
My current research and professional interests lie in Multimodal ML, Medical Imaging, and Graph ML.
Core strengths
Core Strengths & Current Work
Explainable AI (XAI)
Architecting multimodal Concept Bottleneck Models (CBMs) to enhance interpretability and transparency in AI decision-making.
Medical Imaging AI
Have experience working with multiple imaging modalities (MRI, CT, X-ray) and building production-grade vision models for healthcare (Qure.ai). Worked on projects focused on early-detection of chronic diseases (Parkinson's Disease, Lung Cancer, Pneumonia, etc).
Full-Stack AI Products
Designed end-to-end, user-centric AI-based applications, such as SearchIQ (RAG-based search) for employees to chat with company policy documents and knowledge bases. Integrated conversational memory and personalized search suggestions to enhance user experience and engagement.
Open-Source Tooling
Developed accessible, developer-focused repositories, such as Samay (open-source Python library), which unifies training and inference across 10+ time-series foundation models, making it easier for research & developers to adopt and experiment with cutting-edge techniques in time-series analysis.
High-Performance Scaling
Orchestrating distributed PyTorch workflows across HPC clusters to accelerate model training and data processing.
Graphs and Networks
Developed GPLAN , a python-based application for designing 2D-floor plans that optimize for user and design requirements using only graph algorithms and optimization. Led the research and development team involved in adding corridors to floorplans with the power of graph algorithms.
Recognition
Honors & Awards
IEEE BHI 2025 Data Challenge Competition Champion
Oct 2025
Developed an AI/ML-driven depression risk prediction model, winning the IEEE-sponsored competition (backed by NSF & Google)NSF-EMBS-Google Young ProfessionalNextGen Scholar
Sept 2025
Recognized for high-impact Biomedical AI research, including an invitation to present at the IEEE BHI 2025 conference