Quentin Oschatz
ECE PhD Student, CMUSPIRAL Group
About Me
I am a 2nd year PhD student at Carnegie Mellon University in the Electrical and Computer Engineering Department, advised by Prof. Franz Franchetti in the SPIRAL group. My work spans fields such as machine learning, numerical analysis, scientific computing, and code generation.
Before joining the SPIRAL group, I recieved a BSc in Computer Science and Engineering from the Delft University of Technology in the Netherlands, where I also graduated from their Next Generation Robotics Honor's Program. Following that, I completed a MSc in Electrical and Computer Science from Carnegie Mellon University, before being accepted to the PhD program.
Beyond my direct research interests, I am also keenly interested in the world of quantitative finance.
News
Aug 19, 2025
paperOur paper Towards Automated Reasoning Chains for Verification of LLM-Generated Scientific Code was accepted at the 2025 IEEE High-Performance Extreme Computing (HPEC) conference
May 16, 2025
eventAttended the midterm PI meeting for the SOAP project at DARPA in Arlington, VA.
April 29, 2025
eventAttended the AI4Science workshop at Oak Ridge National Laboratories together with my colleague Naifeng Zhang, who presented a brief talk.
Projects
LLM CEREBUS
Cerebus aims to verify the correctness of scientific code, especially when generated by LLMs. Here, mere syntactic correctness is insufficient, as scientific code can easily be plagued with hard-to-diagnose numerical errors despite compiling fine. Leveraging symbolic execution and semantic lifting, we can construct chains-of-reasoning that verify the numerical properties and error bounds of an implementation compared to the mathematical specifications.
SOAP: Scalable On-Array Processing
This DARPA project aims to improve the basic scaling factor behind the processing of certain radar data, specifically when performing beamforming. Here, the clasical O(n3) Minimum Variance Distortionless Response (MVDR) beamforming algorithm is to be improved, ideally to yield a O(n log n) algorithm.
I helped develop and test several ML-based approaches, specifically CNNs acting on complex numbers. Additionally, I built a large part of the radar infrastructure, and helped implement a novel solver algorithm that took advantage of the data structure to reduce the scaling factor to O(n log n).
Publications
HPEC
Towards Automated Reasoning Chains for Verification of LLM-Generated Scientific Code
, Naifeng Zhang, Mike Franusich, Franz Franchetti
IEEE High Performance Extreme Computing, 2025
[Submitted]