Piyushi Manupriya
Interested in Mathematical Guarantees & Elegant Codes

I am an ML postdoc, immensely fortunate to be working with Prof. Anant Raj on optimization & ML theory.
I did my Master's-PhD dual degree from CSE, IIT Hyderabad. My PhD thesis on improving the formulations for Optimal Transport leveraging kernel methods was under Prof. J. Saketha Nath; where I was kindly supported by the Google PhD Fellowship.
During my PhD, I was fortunate to be guided by Prof. Ganesh Ghalme on a Multi-Agent Constrained MAB problem and to be mentored by Prof. M. Vidyasagar (my Doctoral Committee Chairman) and Prof. P.N. Karthik.
During my Master's under Prof. Vineeth N Balasubramanian, I worked on attribution algorithms for interpretable DL.
Prior to that, I did my BTech in CSE from NIT Uttarakhand (in the Garhwal Himalayas), during which I was chosen for the IASc-INSA-NASI Summer Research Fellowship '17 at ISI Kolkata.
I aspire to be an ML theoretician. In future, I expect to also work on problems in online learning, pruning, fairness, privacy and AI for healthcare.
Highlights
I did my Master's-PhD dual degree from CSE, IIT Hyderabad. My PhD thesis on improving the formulations for Optimal Transport leveraging kernel methods was under Prof. J. Saketha Nath; where I was kindly supported by the Google PhD Fellowship.
During my PhD, I was fortunate to be guided by Prof. Ganesh Ghalme on a Multi-Agent Constrained MAB problem and to be mentored by Prof. M. Vidyasagar (my Doctoral Committee Chairman) and Prof. P.N. Karthik.
During my Master's under Prof. Vineeth N Balasubramanian, I worked on attribution algorithms for interpretable DL.
Prior to that, I did my BTech in CSE from NIT Uttarakhand (in the Garhwal Himalayas), during which I was chosen for the IASc-INSA-NASI Summer Research Fellowship '17 at ISI Kolkata.
I aspire to be an ML theoretician. In future, I expect to also work on problems in online learning, pruning, fairness, privacy and AI for healthcare.
- Presenting poster for our AAMAS '25 paper at the ICTS event.
- Attended the BIRS-CMI Game-Theoretic Statistical Inference workshop. Grateful to Prof Shubhada Agrawal & Prof Aaditya Ramdas for letting me join.
- Presented my PhD work on Optimal Transport at IIT Madras. Thanks to Prof. Harish Guruprasad for the kind invitation.
- Our work, Multi-agent Multi-armed Bandits with Minimum Reward Guarantee Fairness, was presented as an Oral at AAMAS '25.
- Presented our TMLR '24 work as an invited poster at ICLR '25.
- PhD Thesis, Improving Optimal Transport using Kernel Methods, available on HAL. BibTeX.
- Defended my thesis (slides) on the eve of Holi, i.e. on my father's (non-Gregorian) birthday.
- Presented our ICML '24 work as an invited poster at ACM ARCS '25.