Aishwarya Mandyam

I am a Computer Science PhD candidate at Stanford University working with Barbara Engelhardt and Emma Brunskill. I'm interested in building statistical/ML tools to improve clinical decision-making.

Outside of research, I enjoy distance running, dogspotting, and eating black forest cake.

Email  /  CV  /  LinkedIn  /  Blog

08/2024: I started an internship at Amazon in NYC, working on reinforcement learning.

04/2024: Adaptive Interventions with User-Defined Goals for Health Behavior Change was accepted to CHIL 2024!

Doubly robust approaches to off-policy evaluation in the prescence of counterfactual annotations.

A modification to Thompson Sampling to enable personalized goal setting using personalized reward functions.

A Bayesian inverse reinforcement learning method that uses conditional kernel density estimation to make computational gains on existing approaches.

A compositional RL framework to find optimal policies in EHR datasets with known heterogeneous treatment effects.

A reinforcement learning guided approach to electrolyte repletion, applied on a cohort from the University of Pennsylvania Medical Center.

An open-source pre-processing and analysis software for MIMIC-IV, a ubiquitous benchmark EHR dataset.

Using machine learning to characterize the relationship between binding aptamers and peptides.

A robust, low density concept for DNA barcoding and storage.

Using influence functions to identify individuals that disproportionately affect the generalization error of the prediction methods used in the Fragile Families Challenge.

Evaluating the results of nanopore sequencing on custom DNA barcodes, and designing new DNA barcodes based on the error analysis results.

Reviewer: NeurIPS 2024, RLC 2024

Organizer: Machine Learning for Healthcare Symposium (ML4H) 2024, Stanford-Berkeley Women’s CS/EE Research Meetup 2024

Made with permission.