Marine Biologist in 2026: Computer Vision in the Reef

Zoe surfaces after a 48-minute reef dive carrying a GoPro, an eDNA sampler, and a passive acoustic recorder. Back on the boat she plugs everything into a ruggedized laptop. By the time the boat reaches the dock, the vision model has counted and ID'd 3,400 individual fish across 71 species, the eDNA lab will return a species list in a week, and the acoustic model has timestamped 17 dolphin clicks and a suspected grouper spawn. Ten years ago she would have hand-counted video for a month.

• Species identification — vision models identify fish, coral, plankton from video.
• Bioacoustics — models detect whales, dolphins, snapping shrimp, fish spawns.
• eDNA analysis — sequencing pipelines with ML classifiers for metabarcoding.
• Satellite remote sensing — chlorophyll, SST, and ice-edge tracking automated.
• Behavior analysis — tracking individual animals across video frames.
• Literature synthesis — grant writing and lit reviews with LLM assistance.

• Tools like FathomNet and VIAME — marine vision model libraries.
• CoralNet — coral cover analysis.
• BirdNET-style acoustic models adapted for cetaceans.
• QGIS with ML plugins — spatial analysis.
• Google Earth Engine for remote sensing at scale.
• Claude and ChatGPT for grant and manuscript drafting.

If you want to be a marine biologist: AP Biology, AP Statistics, physics, learn to swim well and get scuba certified. Undergrad in marine biology, biology, or ecology — take at least one programming class (Python or R). Volunteer on research cruises. A funded PhD is the usual path to research positions; MS is the path to agency, NGO, and consulting work. Learn a GIS stack and one ML framework. Fieldwork still pays in sunburn and seasickness, not in money.