Soroush Sarabi
Summary
Soroush Sarabi — Managing Director of Raderon AI Lab, named inventor on six WIPO PCT international filings, peer-reviewed in MNRAS, PASP and IEEE GRSL, and the principal engineer behind a portfolio of production systems across VR rehabilitation, document intelligence, and Canadian federal-court legal automation.
Skills & capabilities
Experience
Lead applied-AI research and the productisation of computer-vision, document-intelligence and legal-automation systems, and direct the lab's research line in computational neuroscience — neural-network dynamics, attractor states, predictive coding and reinforcement learning (GPT-OSS).
Led AI and data-science work on geospatial 3D data-visualisation and industrial digital-twin pipelines.
Mentored distributed teams building applied machine-learning solutions for AI-for-good challenges.
Co-founded an independent software company; designed and delivered production .NET, WPF and SQL Server systems and database architectures for enterprise and international clients.
Taught programming and information systems and led the programming department; designed curricula and trained hundreds of students and professionals.
Built software and database systems as a software engineer, database designer and IT consultant; earned Microsoft developer certifications (MCITP: Database Developer).
Selected projects
Patents
WO 2020/217082 A1— Generate editable 3D object files from 2D images using deep learning in GAN networks. (2020)WO 2021/198731 A1— AI-based diagnosis of physical characteristics, health and development of agricultural and horticultural plants. (2021)WO 2022/130006 A1— Prognosis, early diagnosis and best-treatment selection via data fusion and AI/machine learning. (2022)WO 2022/064242 A1— Automatic 3D design and reconstruction of constructions and colonies via machine scanning and deep learning. (2022)WO 2021/245439 A2— Machine-vision and AI system to diagnose and reduce risks from animal presence in refineries. (2021)WO 2021/240207 A1— AI, machine-learning and deep-learning prediction, diagnosis and treatment planning for oral and dental diseases in humans and animals. (2021)
Publications
- (2025). Enhanced-HisSegNet · SAR Image Flood Segmentation with Learnable Histogram Layers and Active Contour Model. IEEE GRSL, 22, 1–5. cited by 14
- (2025). Enhancing Skin Cancer Diagnosis using Late DWT and New Swarm-Based Optimizers. ML w/ Apps, 23, 100811. cited by 4
- (2025). Enhanced-HisSegNet: Improved SAR image flood segmentation with learnable histogram layers and active contour model. IEEE Geoscience and Remote Sensing Letters 22, 1-5, 2025. cited by 15
- (2025). Enhancing skin cancer diagnosis using late discrete wavelet transform and new swarm-based optimizers. Machine Learning with Applications, 100811, 2025. cited by 2
- (2024). Enhancing skin cancer diagnosis (scd) using late discrete wavelet transform (dwt) and new swarm-based optimizers. arXiv preprint arXiv:2412.00472, 2024. cited by 5
- (2023). Fundamental Challenges of Neural Network in Handling Sequential Feature of Time Series: NP-Hard Challenge. .
- (2022). Investigation of orbital period & mass relations for W UMa-type contact systems. MNRAS, 510(4), 5315–5329. cited by 37
- (2022). Photometric Study of Six W UMa Systems & Mass–Radius Relations. PASP, 134(1036), 064201. cited by 13
- (2022). Machine learning-based optimization of air-cooled heat sinks. TSEP, 34, 101398. cited by 31
- (2022). ANN Models to Predict HTC and Pressure Drops in Cold Plates with Surface Roughness. ICFFHMT. cited by 6
- (2022). Prediction Accuracy of ANNs in Thermal Management Subject to Network Architectures. MCM '22. cited by 9
- (2022). Artificial neural network to predict pressure drops in heat sinks. Proceedings of the 9th International Conference on Fluid Flow, Heat and Mass …, 2022. cited by 9
- (2022). Application of machine learning to predict thermal performances of heat sinks. Proceedings of the 8th World Congress on Mechanical, Chemical, and Material …, 2022. cited by 6
- (2022). Investigation of the orbital period and mass relations for W UMa-type contact systems. Monthly Notices of the Royal Astronomical Society 510 (4), 5315-5329, 2022. cited by 40
- (2022). The Photometric Study of Six W UMa Systems and Investigation of the Mass–Radius Relations for Contact Binary Stars. Publications of the Astronomical Society of the Pacific 134 (1036), 064201, 2022. cited by 14
- (2022). Prediction accuracy of artificial neural networks in thermal management applications subject to neural network architectures. Proceedings of the 8th World Congress on Mechanical, Chemical, and Material …, 2022. cited by 7
- (2022). Artificial Neural Network Models to Predict Heat Transfer Coefficients and Pressure Drops in Cold Plates with Surface Roughness. Proceedings of the 8th World Congress on Mechanical, Chemical, and Material …, 2022. cited by 4
- (2021). Observational and Theoretical Studies of 27 δ Scuti Stars · Period–Luminosity Relation. PASP, 133(1026), 084201. cited by 33
- (2021). VizieR Online Data Catalog: Times of maxima for 23 delta Scuti stars (Poro+, 2021). VizieR Online Data Catalog 613, J/PASP/133/H4201, 2021.
- (2021). Observational and Theoretical Studies of 27 δ Scuti Stars with Investigation of the Period–Luminosity Relation. Publications of the Astronomical Society of the Pacific 133 (1026), 084201, 2021. cited by 37
- (2021). The First Observational and Theoretical Studies of 27 Scuti Stars with Investigation of the Period-Luminosity Relation. arXiv e-prints, arXiv: 2102.10136, 2021.
- (2020). GENERATE EDITABLE 3D OBJECT FILES ON THE BASIS OF TWO-DIMENSIONAL IMAGES BASED ON DEEP LEARNING AND GAN NETWORK (GENERATIVE ADVERSARIAL NETWORKS). .
- (2019). OC Study of 545 Lunar Occultations from 13 Double Stars. JOE-Journal of Occultation and Eclipse, 2019. cited by 2
- (2018). Evaluating the effect of four unknown parameters included in a latitudinal energy balance model on the habitability of exoplanets. arXiv preprint arXiv:1809.06907, 2018. cited by 3
Certifications
- WIPO PCT — federated GAN for 2D-to-3D reconstruction — World Intellectual Property Organization (2020)
- Neuromatch Academy — Computational Neuroscience — Neuromatch Academy (2024)
Recognition & memberships
NVIDIA Inception Program · AWS Startup · $10K Credit · IEEE Professional Member · ACM Professional Member · Canadian AI Association · CAIAC · BSN Project · contracted ML partner · Microsoft MCITP · Database Developer · Google Scholar · ~163 citations · Omdena · Technical Mentor
Languages
Persian · English