By Emily White
Dr. Dhananjay Bhaskar, Assistant Professor in the Department of Biomedical Engineering, is another new arrival to campus in the RISE AI initiative. Originally from Kanpur, India, Bhaskar grew up in an academic household with chemistry professors for parents. “I grew up in a very academic environment where science and curiosity were just part of everyday life.” This curiosity led him from undergraduate degrees in computer science and math, to Master’s work with Dr. Leah Edelstein-Keshet at the University of British Columbia, and then to a PhD at Brown University with Dr. Ian Wong. While there, he developed methods in topological data analysis and machine learning to solve inverse problems in biological systems.
During his postdoctoral work at Yale, Bhaskar developed geometric deep learning methods for drug discovery, single-cell analysis, and brain dynamics. Bhaskar says, “I’ve always enjoyed art – drawing and painting – and more broadly the intersection of science and art. That naturally pulled me toward studying pattern formation and morphogenesis, and eventually toward learning to recognize patterns via machine learning.”
Bhaskar experiments with different artistic media, such as sketching and painting. He finds it really relaxing while also providing space to reflect and think more deeply. He adds however, “I can be a bit nitpicky when it comes to scientific visualizations and schematics, which I’m sure my students don’t always appreciate.” This intersection of art and science is something Bhaskar utilizes in his research. “During my postdoc, I developed a method to decode mental imagery from brain activity – essentially enabling us to paint pictures from thought.” A lenticular panel showcasing this work is currently exhibited at the Wu Tsai Institute at Yale. We asked Dr. Bhaskar about his research.
Describe your research interests and program.
Broadly, I’m interested in quantitative biology and machine learning, and I work at the intersection of the two in what is increasingly being called mechanistic learning. The idea is to develop biology-informed machine learning models that don’t just make predictions, but can actually help uncover the mechanisms underlying processes like pattern formation, morphogenesis, and gene regulation from biomedical data.
I don’t focus on a specific biological system. Instead, my goal is to develop methods that can capture the structure of biology from high-dimensional, spatiotemporal data across different domains. A big part of my research focuses on understanding the shape and geometry of data – essentially trying to uncover the underlying ‘manifold’ (a useful analogy is the Waddington landscape) on which biological processes actually happen.
At a high level, this gives us a way to move beyond description toward control – to identify the mechanisms driving biological systems, design targeted perturbations, and ultimately develop strategies to shift cells and tissues from disease states toward healthy ones.
How will the RISE AI initiative advance this research?
The RISE AI Initiative has been instrumental in supporting research groups like mine that focus on AI for Science. It has provided access to GPU computing infrastructure that enables us to train and experiment with state-of-the-art agentic, multimodal and foundation models. But infrastructure is just one part of it. There is also strong support for students through training workshops, access to research software engineers to accelerate development, and institutional backing to build and deploy AI systems at scale.
Equally important, RISE-AI has brought together the AI/ML [machine learning] community on campus and created real momentum around team science and interdisciplinary collaboration, especially through connections with RISE-THRIVE, internal funding opportunities and community building events. There’s a lot of energy right now, and it’s been genuinely invigorating to be part of it.
Is there a single person or experience that most influenced your trajectory to where you are today?
My mentor since undergrad, Leah Edelstein-Keshet, has had a lasting influence on both my personal and professional growth – and is someone I still turn to today. She instilled in me a love for lifelong learning, teaching with purpose, and finding joy in work that feels personally meaningful. Most importantly, she taught me to trust my instincts and go after problems that genuinely spark my curiosity.
What opportunities are you excited about?
We’re really in the middle of a renaissance in computational genomics, driven by modern sequencing technologies, single-cell multiome data, and spatial transcriptomics. There’s a huge opportunity right now to develop biologically informed methods to infer cellular trajectories, regulatory networks, and treatment effects from these datasets. At the same time, I’m very interested in inverse design problems – like designing proteins or aptamers – by combining sequence and structural representations. If you’re working in these areas or thinking about related problems, I’d love to connect and explore collaborations. Dr. Bhaskar is excited to collaborate broadly on campus and can be reached at dhananjay.bhaskar@wisc.edu for collaboration requests.