Research & Advising

University of Utah (Incoming Assistant Professor). I am recruiting 2-3 self-motivated PhD students excited about building intelligent data systems at the intersection of databases and large language models.

Advising & expectations

I look for students who are passionate about research! Passion is a must. Most technical skills can be learned with dedication and consistent effort.

Preferred Qualifications

  • Background in computer science, particularly in algorithms and data structures, systems, and databases.
  • Comfortable or familiar with programming (e.g., Python, C++) and frameworks such as CUDA or PyTorch.
  • Prior research experience , publications, and industry experience in related areas are a plus.

Research Topics

We design algorithms and systems for data at scale, with a current emphasis on integrating large language models (LLMs) into database systems. We aim to expand system accessibility to users through natural language interfaces, support multi-model data processing that unifies structured and unstructured data, and explore agentic data systems that reason, plan, and adapt to user intent.

We aim for contributions that are both theoretically grounded and systematically engineered.

Focus areas

NL2SQL & Hybrid Queries(AI SQL)

Robust translation from natural language to SQL/AI SQL with schema grounding, constraint checking, and execution-time validation. Query plans that mix relational operators with LLM calls, including LLM ORDER BY, semantic filters, and reasoning joins. We develop cost models for token-accuracy trade-offs, caching, and verification.

Streaming Algorithms & Sketches

High-performance sketches (e.g., heavy hitters, quantiles, cardinality) with accuracy guarantees, supporting real-time analytics and approximate processing under tight time/memory budgets.

Indexing & Storage

Study indexing structures such as B-trees and LSM-trees, as well as emerging vector indexes (e.g., strong>DiskANN) to support approximate nearest neighbor search. These indexing structures accelerate access to structured, unstructured, and embedding-based data.

Example problems

  • Query optimization for LLM operators: develop cost models and query rewrites to balance accuracy, latency, and token efficiency.
  • Vector-relational fusion: explore ANN and GPU acceleration for data processing, investigating which stages of the data pipeline can be effectively offloaded to GPUs for improved throughput/latency.
  • DB for Agents: investigate how databases should evolve when the primary user is an AI agent — whether to expose more internal details or design abstractions that minimize hallucination and ensure reliability.
  • Streaming Algorithms: design new sketches and explore their integration into ML systems, such as during fine-tuning or inference.

Interested? If one or more of these problems excites you, you’ll likely enjoy working with us.

How to apply / reach out

  1. Apply to the University of Utah. Please submit your graduate application through the official portal and mention my name in your application if you are interested in working with me. Use the official application site: University of Utah Graduate Application.
  2. Email me with the subject line [Prospective Student] and include: (i) a short paragraph on which area(s) you’re most excited about (I’m open to suggestions), (ii) your CV (PDF) and links to any code or papers, and (iii) optionally, a 1–2 paragraph mini‑proposal describing a concrete idea you’d like to explore. Due to volume, I may not be able to reply to every message, but I do read them and will review applications during the admissions cycle.
  3. Current or prospective M.S. students at Utah. If you’d like to join the lab for research or an independent study, email me with your CV, unofficial transcript, and a brief note on the project area you’re interested in.