About

Undergraduate senior at the University of Wisconsin-Madison majoring in Computer Science (with Honors). I have the great fortune of being advised by both Dr. Xiangyao Yu and Dr. Remzi Arpaci-Dusseau.

I am broadly interested in operating systems, distributed systems, databases, and AI/ML systems, with my research interests focusing on how to design systems infrastructure to better facilitate big data and AI/ML workloads. I was previously a Software Engineering Intern on the Lakebase team at Databricks, where I was advised by Dr. Haoyu Huang.

Headshot of Dhruv Aron

Research

UW-Madison Systems Group

  • Investigating how systems design affects local LLM inference and agent performance, context awareness, and persistence.
  • Designing and running experiments to isolate systems-layer performance bottlenecks in local inference engines and open-source agent harnesses.

UW-Madison Databases Group

  • Led a study of the suitability of general-purpose and GPU-native relational databases for genomics and bioinformatics.
  • Designed and ran evaluations comparing DuckDB with BEDtools and SAMtools across runtime, storage usage, and expressiveness.
  • Outlined database extensions and microarchitectures needed to better support these workloads.

Snyder Lab, Stanford University

  • Identified gaps in existing continuous glucose monitoring tools.
  • Developed Glucose360, an open-source Python framework that integrates glucose data with lifestyle events and automates analysis for researchers and clinicians, which is used by professionals at Stanford Health.

Publications

Experience

Databricks

  • Designed and shipped a dynamically resizable PostgreSQL shared-buffer lookup hash table for Lakebase.
  • Reduced the table's memory footprint by 50% and enabled thread-safe dynamic growth and shrinkage while retaining 99% of concurrent query throughput during resizing.
  • Proposed two open-source C patches based on this work, both of which are currently under consideration for official PostgreSQL 20.

TenSixty Biosciences

  • Built a platform for storing, automating, and visualizing computations on .h5ad biological datasets.
  • Combined a Python/Flask API, PostgreSQL, and AWS with a React and D3.js interface for interactive analysis.

Academic Activities

Teaching

Teaching Assistant · CS537 (Undergraduate Operating Systems)

Student Organizations

Coordinator · Undergraduate Projects Lab (UPL)

Organizer · MadHacks 2026

Editor · Journal of Undergraduate Science and Technology (JUST)

Miscellaneous

Outside of research, I enjoy weightlifting and playing soccer and basketball. I also enjoy reading philosophy, mystery, and sci-fi books, as well as watching good movies and television shows.