Ofertas de proyectos

MDS

SERGI ABADAL CAVALLÉ

AC

State-of-the-art models such as LLMs are too large to fit in a single compute node (GPU, NPU, CPU), both for training and inference on a device (e.g., phone, laptop, tablet) or in larger-scale data centers. There is a need to develop optimization techniques to split and place these models onto a distributed set of compute nodes so that the overall system performance is maximized. The research will be focused on optimizing the placement of AI models onto distributed systems considering training time, energy consumption, and computational resources. 

In this thesis, the student will model and simulate distributed AI workloads using both mathematical frameworks and simulators. This encompasses the modeling of network, compute, and memory components within a distributed architecture. Developing new modules to enhance the modeling process. Evaluating and optimizing various parallelization techniques to improve overall system performance.

The Universitat Politècnica de Catalunya · BarcelonaTech offers Master thesis fellowships in the field of LLM training. The research will be supported by Qualcomm and will be carried out in an environment with a strong interaction with leading experts in the field, with opportunities for doing internships in the company.

More information here: https://www.cs.upc.edu/~jordicf/priv/eda/llm_qc.html

Students with strong background on Computer/Data Science and/or Mathematics are required, preferably with skills on:
Algorithms and data structures
Mathematical optimization
Machine and deep learning
Network or Computer Architecture
Oral and written English

MDS

SERGI ABADAL CAVALLÉ

AC

Quantum algorithms are rapidly advancing as a bridge between fundamental mathematics, physics, and emerging quantum technologies. This project portfolio explores three complementary directions at the intersection of quantum linear algebra, circuit design, and graph-based quantum dynamics. By extending classical numerical methods into the quantum domain, developing efficient approaches for generating randomness and expressibility in parameterized circuits, and investigating generalized quantum walks on complex graph structures, the projects aim to deepen our theoretical understanding while advancing near-term applications in computation, communication, and physical modeling. Together, they highlight the diversity of strategies required to harness quantum computing for real-world problems, blending rigorous mathematics with practical implementation challenges.

The candidate will work on one of the following objectives/projects:

1. Quantum Linear Algebra: Quantum QZ Algorithm - Develop a quantum-inspired extension of the QZ algorithm to solve generalized eigenvalue problems, using manifold-based optimization and quantum QR decomposition. Benchmark the approach against classical methods with applications in quantum chemistry, control theory, and machine learning.

2. Quantum Circuit Design: Haar-Random Unitaries and Low-Depth PQCs - Design new parameterizations of the unitary group to enable efficient Haar-random unitary generation. Map these to hardware-friendly circuits and create shallow parameterized quantum circuits that balance expressibility with practical constraints of near-term quantum devices.

3. Quantum Walks: Dynamics on Complex Graphs - Investigate generalized quantum walks on directed, mixed, and embedded graphs, analyzing key dynamical properties and using them for algorithmic applications such as quantum search, Hamiltonian simulation, and secure communication. Explore circuit-based and optical models for near-term realizations.

Background in linear algebra, quantum computing concepts, and numerical methods. Experience with quantum programming frameworks (e.g., Qiskit, Pennylane, or Cirq). Strong analytical and problem-solving abilities, with interest in applications to quantum algorithms, circuit design, or graph-based models.