Who we are
TensorCollective is a Copenhagen-based consultancy founded by four people with backgrounds in data engineering, AI systems, and business strategy. We come from CBS and the University of Copenhagen. We work at the intersection of enterprise data infrastructure and modern AI — building the platforms, pipelines, and agents that teams actually use.
We are trying to become a member of the Anthropic Claude Partner Network, so we can work closely with Claude to help enterprise teams integrate AI into their existing data infrastructure in meaningful, production-ready ways.
The team
Sebastian Søderberg
Data & AI
ss@tensorcollective.io
Sebastian is pursuing a Master's in Business Administration and Data Science at CBS, building on a foundation in Business Administration and Economics. His academic and professional journey is enriched by an international background, having lived and studied in England, Belgium, New Zealand, and Denmark. He focuses on leveraging data analytics, AI, and machine learning to solve real-world business challenges — with a strong drive to develop data-informed strategies that create sustainable business value.
Jeppe Gyland Götzsche
Data & AI
Jeppe is pursuing a Master's in Business Administration and Data Science at Copenhagen Business School, building on a foundation in Business Administration and Mathematical Economics. He has professional experience developing dashboards using Power BI at the Danish Football Association. His interests lie at the intersection of business and data science — driven by using data analytics, AI, and machine learning to address real-world business problems and improve decision-making.
Mads Aslak
Data Engineering & AI Consulting
Mads focuses on AI and agentic frameworks — the infrastructure and patterns that make autonomous systems reliable in production. He is currently developing a multi-agent GitHub automation system using LangGraph and Pydantic AI, with a focus on context engineering, eval frameworks, and observability for multi-step workflows. He also builds MCP servers to extend Claude's capabilities and has explored embedding pipelines for domain-specific image search using CLIP/SigLIP and pgvector.
Johan Wildt
Data & AI
Johan holds a Master's in Business Administration and Data Science from Copenhagen Business School, shaped by international experiences in Hong Kong, Australia, France, and Denmark. His master's thesis — conducted in collaboration with a Danish biopharmaceutical company — explored how transformer-based protein language models (Meta's ESM2) can improve B-cell epitope prediction, evaluating both the scientific and business impact on vaccine design and R&D prioritization. He is driven by the potential of data to power smarter decision-making and sustainable value creation.