A curated collection of my research, organised by track.
Designed SCFM, a training-free safety-alignment method for pretrained T2I diffusion models that steers denoising toward safer generations at inference time via safety-potential-guided rectified flow matching in frozen CLIP embedding space
I worked on some disjoint topics around this broad track.
My undergraduate thesis on generative models for unsupervised speech time-scale modification, under the joint supervision of Prof. Prasanta Kumar Ghosh (SPIRE Lab, IISc) and Prof. Niteesh Sahni (SNIoE) — won 2nd Prize for Best UG Thesis.
I completed 5 PhD-level courses during my undergraduate degree. These are the resources and notes I created while completing them.
From my UG Seminar course, where we were taught how to write reports, papers, and give presentations — these are the resources and notes from that coursework.