⊗

Imagine having a crystal ball able to predict with 99% accuracy what each vessel’s transit will look like before it even leaves the port - the company claiming to have just that is DeepSea Technologies. The Athens-based maritime technology company that brings together a team of artificial intelligence (AI) researchers, engineers, and sector experts is led by Roberto Coustas as its co-founder and CEO. In Coustas's words, Deepsea Technologies is primarily a change in philosophy on how owners and operators manage their vessels. "If I could incapsulate it, it would be a difference between a hunch and experience approach transitioning to a completely data-driven approach of managing and optimising vessels," he tells Splash Extra. In DeepSea’s view, by combing knowledge of a vessel’s DNA with the best weather data, this is the equivalent of waking up each morning and knowing what today will entail - something the company claims weather routing players cannot do. For Coustas, the main problem with weather routing companies is that they make assumptions on how the vessel behaves and, using some theoretical and mathematical models, make their predictions to optimise the vessel. We can be the intermediary to make vessel performance more transparent and better Before co-founding DeepSea with Konstantinos Kyriakopoulos, Coustas worked briefly as a performance engineer and technical coordinator at Danaos, led by his father John. Coming out of Oxford University and seeing what Facebook and Google were doing in terms of machine learning, it seemed absurd how another industry with vessels worth millions of dollars was a bit left behind. Today, he believes we’ve reached the point when shipping will find AI impossible to ignore. "Essentially, AI sounds very fancy, but what it means is that we gather data and create a model of the behaviour of the vessel – we see the strengths and weaknesses of the vessel – one vessel can be very efficient against waves, but it might not behave well under severe wind for whatever reason- it's about utilising the strengths and minimising weaknesses," Coustas explains. The company's models utilise 19 different parameters to have that 99% accuracy, including speed, draught, wind, waves, salinity and water temperature, placing attention on two fundamental things, the data quality, and combining shipping knowledge and engineering knowledge with machine learning. “As we have good data quality, and have gathered all the data and done the underground work, we can create a model that learns the behaviour of the vessel from all of these dimensions and the great thing about all of this is that they are predictable in the future. The salinity of the sea doesn't change from one day to the next, the same as water temperature, underwater currents, and waves three days in advance. It's all of these things that are stable and then we arrive at the crystal ball," Coustas points out. On other maritime AI players, Coustas says: "The primary difference from competition is that DeepSea was founded by people who are machine learning engineers and enthusiasts. Machine learning and AI were at the core of our approach to what we did, and the value we provided was always at the forefront and at the centre of our attention. "At DeepSea we always knew we were going to build a skyscraper and now that skyscraper looks very impressive, but we have spent a long time building a foundation below the ground. We don't have a silver bullet, it's actually a very thorough process of hard work." DeepSea's long-term vision is to be able to bridge the gap between different stakeholders, between charterers and owners. "Instead of working against each other, the goal is to get them to work together. We can be the intermediary to make vessel performance more transparent and better in the end." Without disclosing names, Coustas claims the company has a very large market share in Greece, with clients also in Japan and the US. The impact of DeepSea technology will be disclosed in Q2 this year.

The next best thing to a crystal ball prediction in shipping

DeepSea Technologies claims it can give owners 99% accuracy on any planned transit.

⊗

To continue reading, please login below or subscribe here.

Read Next

Owners advised to cash in

Measuring shipping hubs

April 2024 Review

The on/off risk premium for tankers

Curse of the dry cargo optimist

Owners advised to cash in

Measuring shipping hubs

April 2024 Review

The on/off risk premium for tankers

Curse of the dry cargo optimist

Maritime CEO Forum Singapore

Maritime CEO Forum Monaco

Advertising

Press Releases

Splash

Splash Extra

Splash Special Reports