Alberto Zeni, PhD
Meta
Oslo, Norway
I am currently an AI/HPC Systems Research Scientist at Meta in Oslo, Norway, working on kernels, runtime, collectives, and infrastructure for the Meta Training and Inference Accelerator (MTIA) at data-center scale. Previously, I was a Computer Architect at AMD, where I worked on the design, testing, and validation of the AMD NPU array architecture and was a founding member of the AMD -omics research group.
I hold a Ph.D. in Information Technology and a B.Sc. and M.Sc. in Computer Science and Engineering, all from Politecnico di Milano. My research focuses on high-performance kernels and applications for FPGAs, GPUs, and AI accelerators, with a particular emphasis on genome analysis and, more recently, AI systems and infrastructure at data-center scale. I am also a Co-Founder of GenoGra, a startup delivering scalable, hardware-accelerated genomic analysis software. GenoGra won the StartCup Lombardia 2022 and Switch2Product 2022 competitions, the Chimica Verde Award 2024, and closed a €1M pre-seed round in 2025.
A recent CV.
The best way to contact me is by email: alberto.zeni@polimi.it
selected publications
-
Logan: High-performance gpu-based x-drop long-read alignmentIn 2020 IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2020
-
The importance of being x-drop: High performance genome alignment on reconfigurable hardwareIn 2021 IEEE 29th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM), 2021
-
Optimized implementation of the hpcg benchmark on reconfigurable hardwareIn Euro-Par 2021: Parallel Processing: 27th International Conference on Parallel and Distributed Computing, Lisbon, Portugal, September 1–3, 2021, Proceedings 27, 2021
-
On the Genome Sequence Alignment FPGA Acceleration via KSW2zIn 2023 IEEE International Symposium on Circuits and Systems (ISCAS), 2023
-
TSUNAMI: a GPU implementation of the WFA algorithmIn 2023 32nd International Conference on Parallel Architectures and Compilation Techniques (PACT), 2023
-
Starlight: A kernel optimizer for GPU processingJournal of Parallel and Distributed Computing, 2024
-
Leveraging Difference Recurrence Relations for High-Performance GPU Genome AlignmentIn 2024 33rd International Conference on Parallel Architectures and Compilation Techniques (PACT), Oct 2024
-
MTIA 300: Meta’s First Training Chip with Built-in NICs and Collective Offloading EnginesIn International Symposium on Computer Architecture (ISCA), Jul 2026Alberto Zeni is a member of the MTIA team that published this paper