Welcome!
I am an Assistant Professor in the Department of Computer Science at the Illinois Institute of Technology and the founder and elected chair of the SPEC RG Predictive Data Analytics Working Group.
The overarching goal of my research is to expand the potential of data science in scientific computing by designing robust, efficient, and sustainable system solutions tailored to the evolving needs of data-driven science. As scientific progress increasingly depends on the effective use of data science ecosystems, the diversity of hardware architectures, application demands, and usage patterns poses significant challenges. My work addresses these complexities through a focus on systems and performance engineering, leveraging interdisciplinary expertise to optimize and adapt scientific computing infrastructures for emerging data science applications.
In particular, I see the following key challenges that need to be addressed:
- Resource Optimization: Efficiently allocating and managing resources across diverse hardware platforms to accommodate changing, data-intensive workloads.
- Adaptive Systems: Developing systems that can dynamically adapt to evolving data and model requirements.
- Data Security and Privacy: Safeguarding sensitive data while enabling collaborative data science.
- System-Level Optimization: Optimizing the complex interplay of components within data science ecosystems for maximum performance.
- Sustainable Computing: Minimizing the environmental impact of data science practices.
Most Recent News
| Jul 1, 2026 | Our research paper, “Performance analysis of the homomorphic implementation of K-Means using CKKS”, has been published in Discover Applied Sciences. We present a homomorphic implementation of the K-Means clustering algorithm, showing how cluster comparisons and assignments can be computed on encrypted data. Evaluated with both an offline (bootstrapping) and an online (client re-encryption) approach, both achieve 100% accuracy at a substantial computational cost. |
| May 28, 2026 | Our paper, “Robust Mental Health Detection via Structured LLM Inference: A Study of LangGraph and Chain-of-Verification”, has been accepted in the Research Track of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2026) in September. We study structured inference as a system-level approach to robust mental health detection with LLMs, using a LangGraph-based framework with explicit validation, repair, retry, and abstention to expose failure modes that standard metrics miss. Comparing it against Chain-of-Verification across several models on eRisk anorexia and depression detection tasks, we show that robustness varies substantially across models and cannot be inferred from accuracy alone. |
| Apr 30, 2026 | Our paper, “graphobs: Composable Graph and Observability Queries for Microservice Systems”, has been accepted at the 20th European Conference on Software Architecture (ECSA 2026) in Bolzano, Italy. We introduce a query approach that composes graph and observability queries to support exploration and analysis of microservice systems. |
| Apr 24, 2026 | Our paper, “Conformal LLM Routing with Distribution-Free Safety Guarantees”, has been accepted at the Student Research Workshop of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026). We route queries between a cheaper and an expensive LLM using a conformal gate: a lightweight predictor trained on text embeddings, calibrated with Clopper-Pearson conformal prediction to guarantee that the violation rate among routed queries stays below a target tolerance. To our knowledge, this is the first input-based LLM router with distribution-free safety guarantees. |
| Apr 20, 2026 | Our paper, “Decomposing the Time Series Forecasting Pipeline: A Modular Approach for Time Series Representation, Information Extraction, and Projection”, has been accepted at the 2026 International Joint Conference on Neural Networks (IJCNN 2026). We propose a modular framework that decomposes time series forecasting into representation, information extraction, and projection stages, enabling systematic comparison and composition of pipeline components. |
Selected Publications
- Evaluating Kubernetes Performance for GenAI Inference: From Automatic Speech Recognition to LLM SummarizationIn Proceedings of the 17th ACM/SPEC International Conference on Performance Engineering (ICPE), May 2026ACM Badges: Artifacts Evaluated & Functional
- Evaluation is Key: A Survey on Evaluation Measures for Synthetic Time SeriesJournal of Big Data, May 2024
- The Globus Compute Dataset: An Open Function-as-a-Service Dataset From the Edge to the CloudFuture Generation Computer Systems, Apr 2024
- Methodological Principles for Reproducible Performance Evaluation in Cloud ComputingIEEE Transactions on Software Engineering (TSE), Aug 2021
- Time Series Forecasting for Self-Aware SystemsProceedings of the IEEE, Jul 2020