AI tools, computer vision and applied genomics

Tim Martin Knutsen

Senior scientist at AquaGen AS, Trondheim.

I build practical AI and data tools: LLM-driven workflows, image-based phenotyping, reproducible analysis pipelines, dashboards, and small APIs. Most of this work grew out of applied genomics, where turning messy data into something usable for a real decision is the whole job.

Photo: AquaGen, via NMBU

What I build

The thread running through my work is taking data that is large, noisy, or unstructured and shipping something practical on top of it.

LLM workflows.

Structured extraction, source ingestion, and automated summaries using LLMs and coding agents to turn unstructured text into usable data.

Computer vision.

Image-based phenotyping: extracting reliable biological measurements from images and video, then making the results reproducible.

Reproducible pipelines.

Transparent, automated analysis workflows that other people can rerun and trust, from raw input through to a final result.

Dashboards & APIs.

Lightweight dashboards, FastAPI services, and small PWAs that put analysis and monitoring in front of the people who need it.

Selected projects

A few public examples of the kind of tools I build. These are illustrative side projects, not a complete record of my professional work.

ai-sota-tracker LLM-based extraction over ingested AI sources, with weekly structured summaries. SQLite storage, FastAPI backend, and a PWA frontend.
stock_dashboard Financial and news dashboard with GPT-assisted analysis, pulling market and news data into a single view.
beagle_impute Reproducible genomics workflow: chromosome-wise genotype imputation with Beagle or AlphaImpute2, orchestrated with Snakemake.

Selected Publications

A selection below; see Google Scholar for the fuller, current list. A machine-readable version is also available in publications.json.

Nguyen, D. T., Knutsen, T. M., Sandve, S. R., Lien, S., and Grønvold, L. 2025. Sequence-based chromatin activity modeling and regulatory impact prediction of genetic variants in farmed animals using deep learning. NAR Genomics and Bioinformatics.
Brekke, C., Johnston, S. E., Knutsen, T. M., and Berg, P. 2023. Genetic architecture of individual meiotic crossover rate and distribution in Atlantic Salmon. Scientific Reports.
Brekke, C., and Knutsen, T. M. 2026. Phenotypic sex determines recombination rate and distribution in sex-reversed rainbow trout. Aquaculture.
Knutsen, T. M., Olsen, H. G., Ketto, I. A., Sundsaasen, K. K., Kohler, A., Tafintseva, V., Svendsen, M., Kent, M. P., and Lien, S. 2022. Genetic variants associated with two major bovine milk fatty acids offer opportunities to breed for altered milk fat composition. Genetics Selection Evolution.
Langille, B. L., Kess, T., Nugent, C. M., Messmer, A., Duffy, S. J., Knutsen, T. M., Kent, M., Boyce, D., Gregory, R. S., Gauthier, J., Fairchild, E. A., Pietrak, M., Garcia de Leaniz, C., Consuegra, S., Whittaker, B., Bentzen, P., and Bradbury, I. R. 2024. Trans-Atlantic genomic differentiation and parallel environmental and allelic variation in lumpfish. ICES Journal of Marine Science.

PhD Thesis

Genomics of bovine milk fat composition. Doctoral thesis, Norwegian University of Life Sciences, 2018. Series: PhD Thesis;2018:19.

The thesis examined mutations affecting fatty acid composition in Norwegian Red cattle milk, with emphasis on de novo synthesized fatty acids and the major milk fatty acids palmitic acid and oleic acid.