AI-Music researcher & ML Engineer

Thesis: "Beyond Autonomous Creation in AI Music Generation: Multi-Agent Perception-Policy Separation with Process-Oriented Evaluation for User-Centric, Adaptive Systems"

Thesis: "Neural networks for Music Information Retrieval and algorithmic jazz composition"

Thesis: "Design and Development of E-Commerce Web Applications"

Department of Computer Science and Computational Engineering, UiT - The Arctic University of Norway

• Conducting research on evaluation methodologies for generative music technologies focusing on process-centered frameworks and controlled benchmarking.
• Leading ongoing development and investigation of SCOPE framework through multi-participant study involving music practitioners, music technologists, and researchers.
• Leading the design and development of Sonitra, a modular four-stage pipeline for automatic music transcription assessment that supports digital score generation, rendering, model comparison, and acoustic factor variation.

Department of Music Technology, Norwegian University of Science and Technology (NTNU)

• Conducted research on human-computer interaction in AI music systems, enhancing artistic control in neural audio synthesis
• Collaborated with Prof. Andreas Bergsland to explore user interaction modalities and collaborative music-making environments
• Evaluated AI-driven music generation systems through structured experiments, integrating artist perspectives for improved performance
• Demonstrated application of neural audio synthesis and interactive performance systems

Department of Computer Science and Computational Engineering, UiT - The Arctic University of Norway

• Conducted interdisciplinary research in multi-agent systems and reinforcement learning for music generation
• Developed user-centric AI frameworks for creative applications
• Developed and implemented a representational learning model for unsupervised pattern detection for music analysis
• Established and developed workflow-based evaluation methodology for music generation systems
• Authored and presented peer-reviewed research in conferences and journals
• Led and managed interdisciplinary projects between AI research and artistic practice, ensuring real-world applicability
• Taught 4 graduate and undergraduate courses and supervised over 15 BSc and MSc thesis projects

UiT - The Arctic University of Norway

• Contributed to the Smart Charge project by developing forecasting models for E-mobility, solar energy consumption and market production
• Developed multi-agent systems using zero-intelligence and reinforcement learning to optimize energy market simulations
• Delivered and organized events, lectures, and workshops on machine learning and deep learning techniques

Adaak Financial Institute

• Organized and maintained company's local network infrastructure
• Resolved network issues promptly, enhancing system reliability and user satisfaction
• Conducted routine maintenance and developed tailored network solutions

Talash Argham Institute

• Analyzed and enhanced the local network infrastructure at Talash Argham Institute
• Collaborated with the in-house team to develop tailored solutions for network performance
• Implemented security measures that improved the overall safety of the network environment
• Achieved a more efficient network setup, leading to increased productivity within the organization

Self-employed

• Designed and developed user-friendly websites for diverse clients
• Managed the entire web development process, from requirement gathering to final delivery
• Implemented SEO strategies that improved site visibility
• Gained valuable insights into client expectations, refining project management skills through hands-on experience

IT Orbit Co.

• Supported multiple web development projects at IT Orbit Co., enhancing my technical skills and industry knowledge
• Collaborated with experienced developers, gaining insights into best practices and development standards
• Contributed to the successful completion of projects, improving my problem-solving abilities and teamwork skills

Zed Trading Company

• Led the design team and social marketing department at Zed Trading Company to enhance digital presence
• Managed the development of an e-commerce platform and a streamlined single-page website, improving user experience
• Gained expertise in the Traccar tracking platform, optimizing vehicle tracker modules for operational efficiency

UiT The Arctic University of Norway

MishMash project developing an acoustic benchmarking toolkit that systematically varies reverberation, compression, and distortion conditions to evaluate automatic music transcription tools. The central aim of this project is to distinguish algorithmic weaknesses from recording artifacts. This project is in collaboration with Hans Julius Skaug from University of Bergen, Department of Mathematics.

• Built modular four-stage pipeline—MIDI score → audio rendering via Pedalboard → transcription → evaluation metrics—with interchangeable components for reproducible benchmarking
• Integrated systematic acoustic degradation as explanatory variables to isolate how recording conditions affect transcription accuracy across commercial and open-source AMT models and services
• Evaluating current state-of-the-art automatic music transcription models and services against reference scores using note accuracy, timing, and loudness metrics aligned with MIR community standards

UiT The Arctic University of Norway

MishMash project extending PhD work on SCOPE (Situated Creative Operation and Process Evaluation), a methodology that evaluates AI music tools by examining the entire creative workflow -- from idea formation to iterative revision -- rather than final outputs. It is designed to support multiple stakeholder groups: researchers studying tool impact, developers refining features, and users reflecting on their creative processes. Validated through one-day studio sessions with five professional producers using Suno. This project is in collaboration with Andreas Bergsland from NTNU, Department of Musicology, and Jon Marius Aareskjold-Drecker from UiT, Academy of Music.

• Designed and validated a framework that evaluates AI music tools by their full creative workflow rather than final output, following how producers prompt, revise, and decide to surface where tools help or hinder the process
• Built the evaluation as a feedback loop pairing producers' written reflections with structured scoring, showing how ideas form and evolve while working with AI music platforms
• Ran one-day studio sessions with 5 professional producers of varied ages and backgrounds, who folded Suno-generated material into real compositions and documented their choices, experiments, and points of friction
• Released an open-source toolkit—React frontend, Python backend, Jupyter notebooks, and session templates—so others can apply the framework independently
• Built on a prior year-long PhD study of four open-source systems (MusicGen, Riffusion, Magenta Studio, DDSP-VST)

Digital Xalience AS

Pre-project technical due diligence for AI-powered furniture classification system targeting non-professional users via mobile image capture.

• Validated AI furniture recognition feasibility, assessing foundation models and zero-shot segmentation to inform strategic R&D roadmap
• Evaluated computer vision frameworks including YOLO and Roboflow for enterprise scalability across company sizes
• Assessed dataset availability and quality from Open Images, ShapeNet, Stanford 3D Scanning Repository for training data pipeline

NTNU Music Technology Department

Collaborating with Prof. Andreas Bergsland on human-computer interaction paradigms in AI music systems.

• Investigated latent space representations in neural audio models for enhanced artistic control
• Developing user interaction modalities for enhanced expressive capabilities in neural audio synthesis systems for live electronics
• Advanced understanding of neural audio synthesis applications in live performance contexts

Interdisciplinary Art-AI Project with Artist, Pei-han Lin

Collaborated with interdisciplinary artist Pei-han Lin on a creative AI application for social commentary.

• Developed generative models with StyleGAN by manipulating latent space vectors to control semantic features
• Created a framework for encoding features to explore generative dimensions in synthetic image creation
• Manipulated latent vectors along chosen directional paths with specific boundary conditions for effective generation control
• Integrated AI-generated content into multi-disciplinary installations spanning painting, sculpture, and sound art
• Featured in multiple exhibitions and various art galleries

Interreg-funded EU Collaboration | UiT & Lapland University of Applied Sciences

Developed predictive models for E-mobility, energy consumption and market pricing in Arctic environments.

• Implemented CNN and LSTM neural networks for single-step and multi-step energy load forecasting
• Created multi-agent systems, utilizing zero-intelligence and reinforcement learning for energy markets and V2G (Vehicle-to-Grid) energy optimization
• Incorporated Arctic-specific variables including temperature fluctuations, occupancy patterns, and tourism demand
• Presented findings at CIRED 2023 conference in Rome

Unsupervised anomaly detection for variable-length audio loops using HTS-AT and Deep SVDD.

MIDI segmentation and loop extraction pipeline for symbolic music.

Templates and infrastructure for process-oriented evaluation of music generation systems.

Python implementation of Growing Hierarchical Self-Organizing Maps for unsupervised clustering.

Utilities for GHSOM visualization, analysis, and automatic curriculum extraction.

RL-based combinatorial music generation with hierarchical curriculum and inference-time adaptation.

Dadman, S., Bremdal, B. & Bergsland, A. (2026). Towards Reflective, Situated, and Practitioner-Focused Evaluation and Design Support in Human-AI Music Creation. Submitted to ACM Transactions on Computer-Human Interaction (TOCHI); under peer-review.

Dadman, S. & Bremdal, B. (2026). ARIA: Autonomous Reinforcement-Learning with Intelligent Abstraction for User-Centric Symbolic Music Generation. ResearchGate. doi:10.13140/RG.2.2.22913.31843

Dadman, S., Bremdal, B., Bang, B. & Dalmo, R. (2025). Learning Normal Patterns in Musical Loops. Northern Lights Deep Learning Conference. https://openreview.net/forum?id=Pr22XVnMW1

Dadman, S., Bremdal, B. & Bergsland, A. (2025). Workflow-Based Evaluation of Music Generation Systems: Open-Source Case Study. arXiv. doi:https://doi.org/10.48550/arXiv.2507.01022

Dadman, S. & Bremdal, B. (2024). Crafting Creative Melodies: A User-Centric Approach for Symbolic Music Generation. Electronics. doi:10.3390/electronics13061116

Dadman, S. (2023). Boosting Creativity with AI: Exploring Advanced Models, Multi-Agent Systems, and Design Grammar. . doi:10.13140/RG.2.2.24877.67041

Dadman, S. & Bremdal, B. (2023). Multi-Agent Reinforcement Learning for Structured Symbolic Music Generation. Advances in Practical Applications of Agents, Multi-Agent Systems, and Cognitive Mimetics. The PAAMS Collection. doi:10.1007/978-3-031-37616-0_5

Naeimaei, A. & Dadman, S. (2023). Mindful Integration of AI in the Design Industry: Opportunities and Implications. NORA Annual Conference.

Dadman, S. & Bremdal, B. (2023). Using Light Weight Electric Vehicles for V2G Services in the Arctic. IET Conference Proceedings. doi:10.1049/icp.2023.1230

Bremdal, B. & Dadman, S. (2023). Predicting Peak Prices in the Current Day-Ahead Market. IET Conference Proceedings. doi:10.1049/icp.2023.1244

Bremdal, B., Ilieva, I., Tangrand, K. & Dadman, S. (2023). E-Mobility and Batteries—A Business Case for Flexibility in the Arctic Region. World Electric Vehicle Journal. doi:10.3390/wevj14030061

Dadman, S., Bremdal, B., Bang, B. & Dalmo, R. (2022). Toward Interactive Music Generation: A Position Paper. IEEE Access. doi:10.1109/ACCESS.2022.3225689

Dadman, S., Bremdal, B. & Tangrand, K. (2021). The Role of Electric Snowmobiles and Rooftop Energy Production in the Arctic: The Case of Longyearbyen. J. Clean Energy Technol.

poster Learning Normal Patterns in Musical Loops — Northern Lights Deep Learning Conference 2026 (Tromsø, NO) [paper]

lecture Control and Explore: Neural Audio Synthesis with VAEs for Live-Electronics and Interactive Performances — Faglig Forum, Music Technology Department, NTNU, Norway (Trondheim, NO) [slides]

lecture Artificial Intelligence and Music: Deep Learning and Agents for Music Generation — SINTEF-ZEB Lab, Trondheim, Norway (Trondheim, NO) [slides] [video]

podcast A user-centric approach for symbolic music generation — CreateMe podcast, University of Agder (Agder, NO) [audio]

lecture Melody and Machine: Exploring AI's Role in Music Creation — Algoritmi, UiT The Arctic University of Norway (Tromso, NO) [slides] [video]

poster Integration and influence of artificial intelligence in the design industry — NORA Annual Conference 2023 (Oslo, NO) [paper]

demo Application of multi-agent systems and reinforcement learning methods in interactive music generation — AI+ Conference (Oslo, NO) [video] [code]

lecture Interactive music generation with Artificial Intelligence — University of Oslo, Brain Talk webinar (Oslo, NO) [video]

lecture Application of multi-agent systems and reinforcement learning methods in computational creativity and music generation — UiT, The Arctic University of Norway, Bodo (Bodo, NO) [slides]

exhibition Art x AI — UiT, The Arctic University of Norway, Narvik (Narvik, NO) [video]

exhibition If I Were Standing in your Shoes — Tromsø kunstforening (Tromsø, NO) [article]

exhibition This is a Protest Gesture to Showcase Norway's Violation of Basic Human Rights — Storgata (Tromsø, NO) [article]

presentation Application of deep learning methods in music generation — UiT, The Arctic University of Norway, Narvik PhD. Conference (Narvik, NO) [slides] [paper]

demo Application of deep learning methods in symbolic music generation — AI+ Conference (Oslo, NO) [video] [code]

demo Deep Jazz Composer — UiT, The Arctic University of Norway, Narvik Research Week (Narvik, NO) [video] [code]

interview Composing Jazz with Deep Learning — NRK P3 and Nyheter (Norway) [audio]

media Har utviklet kunstig intelligens som lager jazzmusikk — Forskning.no (Oslo, NO) [article]

media Utvikler kunstig intelligens som komponerer jazzmusikk — UiT Highlights (Oslo, NO) [article]

media Narvik-studenten utvikler kunstig intelligens som komponerer musikk — Fremover.no (Narvik, NO) [article]

media Musikk i koder — Jurnalen.oslomet.no (Oslo, NO) [article]