Cristiano Malossi

I am a Principal Research Scientist, AI Manager, and Global Research Lead for AI for Partial Differential Equations (PDEs) at IBM Research.

Since mid-2025, I have been leading an organization focused on developing next-generation AI algorithms that deliver high-fidelity physical predictions at a fraction of the computational cost of traditional numerical simulations. As part of this effort, we developed a novel Neural Operator model capable of accurately modeling turbulent fluid dynamics at high Reynolds numbers over complex geometries. The model achieves ~1% error compared to state-of-the-art Computational Fluid Dynamics (CFD) solvers, while reducing time-to-solution from hours to seconds.

Previously, I led innovation for IBM products in computer vision. In 2023, I founded the Visual Prompting Lab, a cloud service that enables users to build computer vision models with minimal effort. Earlier, in 2020, I launched Inspecto, a cloud service for large-scale asset inspection. Both services achieved strong client adoption and were successfully integrated into the IBM Maximo portfolio, strengthening IBM’s enterprise AI capabilities.

From 2017 to 2019, I led global research on Neural Architecture Search (NAS). In 2018, my team released NeuNetS, the first IBM Cloud engine for neural network synthesis for both image and text classification. This work resulted in multiple publications at leading AI conferences, including AAAI and NeurIPS.

Earlier in my career at IBM, I focused on novel computing algorithms that leverage approximation and automation to improve energy efficiency. This research produced multiple patents and publications and led to a major EU grant for OPRECOMP, a project I personally coordinated from 2017 to 2020.

In 2015, I received the ACM Gordon Bell Prize for scaling an implicit PDE solver to 1.5 million cores with 97% parallel efficiency. I was the first IBM author on the paper and led code optimization efforts for scalability and performance, including MPI optimization and SIMD vectorization. I am an ACM Senior Member and Distinguished Speaker, and I have served on the technical program committees of major international conferences, including SC, ISC, IPDPS, AAAI, NeurIPS, ICML, and DATE.

I earned a PhD in Applied Mathematics from EPFL, where my thesis on parallel algorithms and numerical methods for cardiovascular simulations won the IBM Research Prize for Scientific Computing. I previously graduated summa cum laude from Politecnico di Milano with a B.Sc. in Aerospace Engineering and an M.Sc. in Aeronautical Engineering.

Honors and Awards:
2024 - "IBM Tech" Recognition Program for Top 1% IBM Technical Contributors in 2023
2023 - ACM Senior Member
2022 - MICAD Best Paper Award
2022 - Best Paper Award at CVCIE @ ECCV2022 workshop
2019 & 2022 & 2025 - ACM Distinguished Speaker
2016 - IEEE/ACM IPDPS Best Paper Award
2015 - IBM Pat Goldberg Memorial Best Paper Award
2015 - ACM Gordon Bell Prize
2013 - IBM Research Prize for Computational Science (for the PhD thesis)

Code and Tools:

• NeuNetS: Neural Network Synthesizer (IBM Cloud)
• OPRECOMP: EU Project on Transprecision Computing (GitHub)
• BAGAN: Keras implementation of BAlancing GAN (IBM GitHub)
• IBM Optimized High Performance Conjugate Gradient (IBM GitHub)
• LifeV: Library for the numerical solution of PDEs with FEM (BitBucket)

Videos:

• Earth Mantle Flow Simulation (IBM Research Youtube channel)
• Visual Prompting Lab, Inspecto, Blood Flow and Basin Simulations (My Youtube channel)