Viapontica AI

Please describe your company and your position there.

I am the COO of a quantitative foresight company based in France.

For what projects/services did your company hire Viapontica AI, and what were your goals?

We were developing a POC tool for tracking, analysing and assessing an emerging asset class in the virtual economy. We hired Viapontica AI to provide data engineering services for the project, including data ingestion and pre-processing, data transformation, data enrichment, data analytics, and creating flexible data outputs to our desired dashboard.

How did you select this vendor and what were the deciding factors?

We found Viapontica AI through a referral and assessed them against a couple of other proposals. We hired Viapontica AI because they demonstrated a deep understanding of our problem; flexibility to work on an experimental and open-ended problem set; and contributed highly insightful ideas on how to augment our approach early on in the proposal process.

Describe the scope of work in detail, including the project steps, key deliverables, and technologies used.

Viapontica AI set up our cloud environment and tools, including Google Cloud Platform/BigQuery, and Tableau, and built a full data processing pipeline for us.

They also defined the ML requirements, created the ML datasets, and worked with us on developing metrics to be extracted from the data, as well as created preliminary metric visualisations in the Tableau dashboards.

They also provided technical assessment and early exploration of a new transaction data API related to the asset class we were studying, and compared it with another high-priority data source.

How many people from the vendor's team worked with you, and what were their positions?

There were 2 people in the core team, including a data engineer/data scientist and a managing director.

Can you share any measurable outcomes of the project or general feedback about the deliverables?

Viapontica AI's work freed up our data scientists to focus on data modelling and accelerated our project by 3-4 months.

Describe their project management style, including communication tools and timeliness.

Viapontica AI provided effective, predictable and adaptive project management via Slack, email, ClickUp and Google Drive. They provided best-in-class documentation throughout the project and created an efficient operating rhythm.

What did you find most impressive or unique about this company?

Viapontica AI provided technical expertise and contributed intellectually to the project by participating actively in developing metrics and visualisations, and planning the future stages of the work. They were very responsive and adaptable to the open-ended exploration at the heart of the project.

Are there any areas for improvement or something they could have done differently?

None.

Please describe your company and your position there.

I am a Technical Director of an international built environment consultant.

For what projects/services did your company hire Viapontica AI, and what were your goals?

We teamed with Viapontica for a R&D project to develop automation in processes for tunnel examinations. Viapontica's role was to develop data processing, machine learning and presentation of results. The goal was to develop tools to automate the detection of defects in rail tunnels.

How did you select this vendor and what were the deciding factors?

We were introduced to Viapontica at a Networking event held by the project sponsor and recognised their strong relevant experience in using ML techniques to detect similar defects in other applications.

Describe the scope of work in detail, including the project steps, key deliverables, and technologies used.

We provided raw data of different types which Viapontica then used to develop processes to clean and prepare the for machine learning which they also developed for the types of source data (spatial data and image data) and the type of defects to be identified. The final step was presentation of the results, for which they developed a prototype App.

How many people from the vendor's team worked with you, and what were their positions?

I worked directly with Vesko (Director) for the most part, although some on our team were in direct contact with other Viapontica team members while annotating data for the Machine Learning process.

Can you share any measurable outcomes of the project or general feedback about the deliverables?

The project was a development project and, as such, outcomes can be uncertain and in this case there was a chain from data collection through to processing and final results which was under development. Some of the input data was sub-optimal which limited Viapontica's outputs however, their developments were in line with expectations.

Describe their project management style, including communication tools and timeliness.

Viapontica exhibited a flexible approach to project management, adapting to upstream constraints to apply the required resource at the right time. The project began just as the first Covid lockdown came into effect so collaboration was all through Teams and email.

What did you find most impressive or unique about this company?

Vesko and his team came across as very knowledgeable about their field in Machine Learning, selecting and adapting a range of techniques to tackle the problem. There was also a thirst for knowledge and understanding about the sector they were applying their methods to, i.e. civil engineering structure examinations.

Are there any areas for improvement or something they could have done differently?

Overall, we were satisfied with the way things went on the project. Outcomes could have been improved with more face to face workshops but COVID restrictions at key stages in the project made this difficult. Improvements to the upstream data collection (not in Viapontica's control) could also improve outcomes.

Please describe your company and your position there.

Creative Informatics, is an ambitious research and development programme based in Edinburgh, which aims to bring the city’s world-class creative industries and tech sector together, providing funding and development opportunities that enable creative individuals and organisations to explore how data can be used to drive ground-breaking new products, businesses and experiences. As programme manager I oversee delivery of the programme including management of the Creative Informatics delivery team and leading our programme of R&D funding calls.

For what projects/services did your company hire Viapontica AI, and what were your goals?

Following an open funding call, Viapontica were successfully selected to respond to a Challenge project proposed by The List, work which was funded by Creative Informatics as part of our Challenge programme. The List is an Edinburgh-based events listings magazine and data company, and they work with a range of third party data clients, who often request multiple images for each event. While it is straightforward for a computer to crop an image to a particular size, it’s important that the resulting image still looks good, and preserves the overall impact of the original. Elements in the image, in particular people, should not be cropped in half. Being able to provide the maximum number of images possible in a range of formats that fits every single app, website, printed publication (and more) is very hard. Viapontica identified a technological solution to this problem which they developed into a solution as part of an R&D project. For more details of the Challenge, our public call can be viewed at: https://creativeinformatics.org/challenge/the-list/

How did you select this vendor and what were the deciding factors?

Viapontica were selected through an open call for responses. A number of other companies proposed solutions but Viapontica's response stood out for two main reasons: they understood the technical complexity of what can appear to be a relatively straightforward challenge (a classic case of something it is easy for humans to do, but very hard for computers to do well); and they demonstrated substantial expertise in AI and particularly techniques for processing and analysing images, including experience from other sectors (at that time particularly around the condition of highways), which made it clear that they would be able to deliver the work they had proposed.

Describe the scope of work in detail, including the project steps, key deliverables, and technologies used.

Viapontica worked with The List to understand their current processes and technologies, as well as their data sets. They researched appropriate AI approaches, working with training data to understand the relevance to the kinds of events images The List works with. Viapontica delivered 'Cream of the Crop', a service based on a very clever and effective AI pipeline to handle key image elements and identify the focal point of images to enable smart cropping, as well as a mechanism for outputting images in a multitude of sizes and ratios.

How many people from the vendor's team worked with you, and what were their positions?

We worked mainly with two members of Viapontica: The Director (Vesko) and one of their Machine Learning Engineers (Vera).

Can you share any measurable outcomes of the project or general feedback about the deliverables?

Viapontica delivered a high quality MVP (Minimum Viable Product) that absolutely met the goals of the project - including not only image cropping but sophisticated management of text in the images (common in the data set at hand). The solution saves significant manual effort and the trial results seen were excellent. They also looked at emerging AI processes which could be used to further develop the solution in the future to further enhance the cropping and image editing potential, which was exciting to see. We were very happy with the outcome of the project as the funder, and we understand that The List were also very satisfied with the work undertaken and the knowledge shared and gained in the process.

Describe their project management style, including communication tools and timeliness.

Viapontica were organised and communicated very clearly. Timelines were met (or only changed as a result of external dependencies) and they maintained good relationships with ourselves and The List in the process.

What did you find most impressive or unique about this company?

Viapontica clearly understand AI in depth and we found their enthusiasm for the project, and additional exploration of newer techniques and opportunities particularly impressive. I would also say that Viapontica do an exceptional job at communicating complex and new technologies and approaches very clearly - something which is hard to do and very rare in the AI space.

Are there any areas for improvement or something they could have done differently?

We were very happy with the work undertaken. I think we would like to see how the service developed can be taken further or opened up to other clients - but that is more of an aspiration rather than a flaw in their delivery as this project was always designed to be an intense R&D process towards an MVP. They were very ambitious and we were very pleased with what they were able to deliver.

Introduce your business and what you do there.

I work for the UK National Archives as a product manager in the digital archiving department.

What challenge were you trying to address with Viapontica AI?

We're at the beginning of developing an extra service for digital records for our government users to access their digital data that have been transferred to us, so we enlisted their help to guide us through the process.

What was the scope of their involvement?

Viapontica AI worked with us on our discovery and alpha phase for website development. Primarily, it was a lot of user research and developing prototypes to test with our government users to build the service. They didn't do coding, but they've given us technicals or blueprints and developed a process that will become the background for our online service.

What is the team composition?

We worked directly with four members of their team, including their project manager, a user researcher, a technical lead, and a UX designer.

How did you come to work with Viapontica AI?

We published an outcome on the government digital marketplace that they applied for. We chose them because they gave a very strong application in terms of ability and ways of working.

What is the status of the engagement?

We worked together from October 2021–March 2022.

What evidence can you share to demonstrate the impact of the engagement?

We've just finished this project and had a service assessment of our alpha project focused on how we look and what we've achieved, and so far, the results of their work have been very positive.

How did Viapontica AI perform from a project management standpoint?

They were excellent at meeting deadlines and communicating with us. We worked agile and had regular meetings that we attended. Regarding tools and communication, we used Slack and Trello.

What did you find most impressive about them?

Their team was very keen to understand the kind of product they were working with and fed that into their work.

Are there any areas they could improve?

I don't think there's anything they can improve on.

Any advice for potential customers?

Try as much as possible to be clear with communication regarding the details of your project, and if possible, have face-to-face meetings with them.