Formation
Schools
Workshops
Funding possibilities
Job Offers
PHINDER: high-resolution optical sensing
Date of posting: Feb 1st 2026
Deadline: 31st March 2026
Target: Degree in physics, electronics engineering or in a closely related discipline, or equivalent degree.
Key words: Photonic sensors, neuromorphic computing.
Project: PHINDER: high-resolution optical sensing.
Duration: 4 years.
Job description:
The HEP group at University of Oviedo (Spain) is looking for a PhD student to join our team and work on the “PHINDER” EIC Pathfinder Open project. The PHINDER consortium aims to revolutionize high-resolution optical sensing by combining photonic sensors with spiking neural networks to achieve picosecond-scale event detection and analysis, dramatically improving energy efficiency and spatiotemporal resolution over current technology. The Oviedo HEP group plays a leading role in the MODE Collaboration (https://mode-collaboration.github.io/ ) on energy-efficient experiment design, and in several activities of the CMS Collaboration (e.g. Machine Learning Group, Phase-2 upgrade work for the DT and L1 trigger upgrade and several physics analyses).
Candidates should have, or be expected to obtain by the call deadline, a master’s degree in physics, electronics engineering or in a closely related discipline, or equivalent degree, be highly motivated to join a multidisciplinary team, and excited to work in an international collaborative environment. Previous experience in data science, machine learning applications, pattern recognition algorithms, spiking networks, or neuromorphic computing will be considered a plus.
The selected candidate will join the HEP group at University of Oviedo, and the PHINDER project with the goal of designing and optimising surrogate models and spiking neural models to be implemented in neuromorphic hardware manufactured by the consortium and co-designing said hardware. The position is initially for 4 years (tentative starting date in April/May, subject to administrative constraints) until completion of the PhD. The candidate is expected to be based in Oviedo with stays at other PHINDER partners (Sweden, Italy, Netherlands, Belgium, Spain) in case they are needed.
Place: Oviedo, Spain.
Contact: Prof. Pietro Vischia, E-mail: vischia@uniovi.es
Interested candidates should submit via email the following documentation:
1. Motivation letter;
2. Curriculum vitae;
3. Arrange for at least two letters of reference to be sent to the same email address.
Note: Review of complete applications will begin on February 15th.
Scalable Ultra-fast Nowcasting with Real-time In-situ Embedded hardware (SUNRISE)
Date of posting: Spring 2026
Target: MsC students in physics, electronic engineering, telecommunications engineering, computer science or similar.
Key words: Artificial Intelligence (AI), Solar nowcasting; Edge AI on FPGA; Real‑time PV forecasting; Intra‑minute cloud prediction; Grid stability and penalty reduction.
Project: Scalable Ultra-fast Nowcasting with Real-time In-situ Embedded hardware (SUNRISE)
Description: SUNRISE is a technology‑transfer project that applies state‑of‑the‑art machine‑learning‑on‑FPGA expertise, originally developed for high‑energy‑physics triggers, to ultra‑fast, on‑site solar power nowcasting, with the aim of improving grid stability, reducing penalties and lowering the carbon footprint of forecasting. The SUNRISE project will develop and validate a real‑time nowcasting system for photovoltaic (PV) plants, based on AI models deployed directly on customizable hardware (FPGAs, SoCs, SoMs). All‑sky cameras and environmental sensors will feed spatio‑temporal ML models capable of predicting intra‑minute cloud motion and PV power output with sub‑second latency, eliminating dependence on remote data centres. The work plan is structured in four work packages, progressing from model development and compression, through translation to firmware and deployment on COTS FPGA/SoC platforms, to a controlled‑environment demonstrator and finally a field prototype in collaboration with industrial partners, targeting >95% accuracy for intra‑minute forecasts, 20–30% reduction in grid penalties and TRL 6–7 for market‑ready validation.
Duration: 2 years
Job description: We are seeking a researcher with great motivation and some background in machine learning and computer vision, with experience in spatio‑temporal modelling and an interest in deploying models on resource‑constrained hardware for real‑time applications. The main duties include designing, training and benchmarking ML models such as convolutional/recurrent architectures, ConvLSTMs and transformer‑based approaches for cloud‑cover and PV power nowcasting using all‑sky images, meteorological information and plant data, with performance targets of at least 90% accuracy on historical data and 95% in controlled tests. The researcher will develop and maintain robust data pipelines, encompassing camera calibration and fisheye correction, cloud segmentation, synchronization of sensors and PV measurements, dataset preparation and systematic comparison with existing nowcasting solutions. A key part of the role is performing model compression and optimization – including quantization, pruning and knowledge distillation – to ensure compatibility with FPGA/SoC deployment in close collaboration with the firmware team, while preserving accuracy and low latency.
Place: Universidad de Oviedo, Spain
Contact: folguerassantiago@uniovi.es, fjcuevas@uniovi.es, fernandezmenjavier@uniovi.es.
Monte Carlo acceleration with FPGAs
Date of posting: Spring 2026
Target: MsC students in physics, electronic engineering, telecommunications engineering, computer science or similar.
Key words: MC simulation, Acceleration, FPGAs
Project: Monte Carlo acceleration with FPGAs
Description: We are building ultra-fast, ultra-green Monte Carlo engines on next-generation FPGAs and edge devices, turning today’s CPU-bound simulations into real-time tools.
Starting from MadGraph in high-energy physics and extending to finance and engineering, we target order-of-magnitude gains in speed and performance-per-watt.
Our FPGA designs exploit deep pipelining, massive concurrency, and precision-tuned arithmetic to squeeze every joule of useful computation from the hardware.
A systematic benchmark campaign against highly optimized CPU and GPU codes will quantify runtime, energy, and scalability on realistic workloads.
By demonstrating MC acceleration from data-center cards down to edge SoCs, we aim to make HPC-grade simulations available anywhere, not just in big clusters.
The outcome will be open, reusable IP blocks, software hooks, and best practices that lower the barrier to MC-on-FPGA adoption across multiple disciplines.
Duration: 1 year
Job description: We seek an engineer or physicist to design and optimize FPGA-based accelerators for Monte Carlo simulations in high-energy physics, finance, and engineering.
You will develop C/C++ and HLS (e.g. Vitis HLS) kernels for modern FPGA platforms (Alveo/Versal and Zynq-class SoCs), focusing on deeply pipelined, parallel architectures.
Your responsibilities include implementing parallel random-number generators, floating/fixed-point datapaths, and resource-aware arithmetic units.
You will profile and benchmark FPGA designs against optimized CPU and GPU implementations, including power and energy measurements.
You will integrate the accelerators into existing Monte Carlo frameworks (starting with MadGraph), writing host-side software and testing infrastructure.
You will collaborate closely with HEP and HPC experts to refine algorithms, document designs, and contribute to publications and open-source releases.
Experience with FPGA toolchains (Vivado/Vitis or similar), hardware–software co-design, and parallel programming is expected; familiarity with MC methods is a strong plus.
Place: Universidad de Oviedo, Spain
Contact: folguerassantiago@uniovi.es, fjcuevas@uniovi.es, fernandezmenjavier@uniovi.es.
Optimization of industrial processes using FPGA and ML algorithms
Date of posting: Spring 2026
Target: MsC students in physics, electronic engineering, telecommunications engineering, computer science or similar.
Key words: Simulation, Industrial processes, AI, Acceleration, FPGAs
Project: Optimization of industrial processes using FPGA and ML algorithms.
Description: This project aims to boost the efficiency and sustainability of industrial production lines by combining FPGA-based hardware acceleration with advanced machine-learning algorithms deployed close to the process.
We will design FPGA architectures that ingest sensor and control data in real time, executing ML models for anomaly detection, predictive maintenance, and adaptive control with minimal latency and power consumption.
By tailoring the hardware to the specific workloads and co-designing the ML models for embedded execution, the project targets significant gains in throughput, energy efficiency, and overall equipment effectiveness.
Validated on representative industrial use-cases, the resulting platform will provide a reusable blueprint for smart factories that require reliable, low-power, real-time optimization at the edge.
Duration: 1 year
Job description: We are looking for a researcher/engineer to lead the development of FPGA-accelerated, ML-driven optimization solutions for industrial processes.
The candidate will design and implement real-time data-processing and control pipelines on modern FPGA platforms (e.g. Zynq/Versal/Alveo), integrating sensor inputs, pre-processing, and ML inference under tight latency and power constraints.
They will co-design machine-learning models (for anomaly detection, predictive maintenance, and adaptive control) with the hardware architecture, including model compression, quantization, and streaming interfaces for efficient embedded execution.
Responsibilities include benchmarking against CPU/GPU baselines, profiling performance and energy consumption, and iterating towards maximum throughput and performance-per-watt on realistic industrial workloads.
The candidate will collaborate with domain experts from industry to understand process KPIs, define representative use-cases, and validate the solution on real or emulated production lines.
Strong experience with FPGA design flows (Vivado/Vitis HLS or similar), C/C++, hardware–software co-design, and basic control/ML methods is required; familiarity with edge-computing or industrial automation is a strong plus.
Place: Universidad de Oviedo, Spain.
Contact: folguerassantiago@uniovi.es, fjcuevas@uniovi.es, fernandezmenjavier@uniovi.es.
Enhancing the LHCb experiment at CERN with sustainable technology
Date of posting: 1st October 2025
Funding source: CSIC - Programa de Atracción y Retención de Talento Digital [MMT24-IFIC-3]
Target: MsC students in physics, computing science or similar.
Key words: Artificial Intelligence (AI), Technologies for processing big amounts of data and information, High performance computing, Green algorithms.
Project: Enhancing the LHCb experiment at CERN with sustainable technology
Description: The LHCb collaboration at CERN is currently using a pioneer system of data filtering in the trigger system, based on real time particle reconstruction in Graphics Processing Units (GPUs). This corresponds to processing 5 TB/s data rate of information and has required a huge amount of hardware and software developments. In particular, the Allen project consists of more than 365 algorithms which executes real-time vertex finding and reconstruction, fast track particle reconstruction, calorimeter clustering and muon identification with very high efficiency and high throughput. In several of its algorithms Allen uses a fast and powerful Neural Network (NN) to suppress reconstructed objects from random hits in the detectors. It allows the suppression of false trigger selection decisions which contribute to the background. An important feature to take into account is the corresponding power consumption and sustainability of this framework. This implies the development of tools to optimize energy usage through efficient computing architectures and algorithms. It involves exploring hybrid computing platforms to take advantage of the strengths of various hardware systems in versatility, parallelization, and customizability, alongside implementing efficient software solutions for both, increasing the physics output and reducing the energy consumption.
Duration: 3 years
Place: IFIC - Valencia (CSIC/ U. Valencia)
Contact: Arantza.Oyanguren@ific.uv.es, Alvaro.Fernandez@ific.uv.es
