Who they are
DefinedCrowd is a start‑up with a platform that enables data scientists to collect, structure, and enrich high‑quality data used to train artificial intelligence algorithms. Because data quality and processing speed are fundamental to its business, the company relies on Colt for fast, efficient connectivity to access all its data.
Ricardo Gusmão, IT & Information Security Manager at DefinedCrowd, highlights Colt’s international presence as a decisive advantage. He sees Colt not just as a supplier, but as a partner supporting the company through its growth phase.
The Business Drivers
Artificial intelligence depends on high‑quality training data — an area often overlooked. DefinedCrowd was founded in 2015 to transform unstructured training data into structured, reliable datasets. The company works with large volumes of cloud‑based data, requiring high‑speed, stable connectivity, which Colt provides through IP Access.
Clients can either supply their own data for structuring or ask DefinedCrowd to collect it. Because data scientists spend up to 80% of their time preparing data, the platform is designed to free them from this heavy workload.
"Clients go to the platform and can customise the type of data they want. Let’s say they are working on a voice recognition system and they need training data. They decide how many scripts they want people to read, how many times they want them read and in what languages. Afterwards, these requirements are transformed into microtasks which are funnelled to our human-in-the-loop community." - Catarina Salteiro, Global PR & Communication Manager at DefinedCrowd
How the platforms works
DefinedCrowd’s human‑in‑the‑loop community includes 130,000+ contributors across 70+ languages and dialects, performing microtasks such as tagging images and annotating text. Combined with the company’s engineering expertise, this supports advanced applications in voice, NLP, and computer vision.
“Our community does the microtasks, which enables us to get high-quality data sets. At this point, we have highly qualified members to provide the data our clients need, which allow us to guarantee 95 to 98% of quality,” says Salteiro.
To run this infrastructure, DefinedCrowd uses Colt’s symmetrical 1 Gbps IP Access. Colt also provides crucial web security services, strong SLA conditions, and highly reliable availability.
"Availability is very important, especially in companies like ours, who depend on communication. This availability, as well as reliability, are the most important factors, and Colt is capable of providing them”. “I really see Colt as a partner, not just as a supplier." - Ricardo Gusmão, IT & Information Security Manager
Growth and Global Expansion
DefinedCrowd supports clients across industries including healthcare, automotive, energy, fintech, retail, and media. Many are Fortune 500 companies, such as BMW, Mastercard, and Yahoo Japan. Amazon is a key investor and recommends DefinedCrowd based on its work supporting Alexa. The company is also a Microsoft co‑sell partner, with its solution integrated into IBM Watson.
Following an $11.8M USD investment round in 2018, the company has rapidly scaled, operating across Lisbon, Porto, Seattle, and Tokyo — all cities where Colt provides fibre connectivity, enabling a single global provider.
To support future expansion, DefinedCrowd is considering adopting a dedicated Microsoft Azure cloud environment.
“This is why having Colt present in the countries where we operate is so important, because the information access speed is extremely important to us”, says Gusmão. “Being located in places where we have offices can benefit us when dealing with connections”. These are the foundations for the goal the start-up has set for themselves: to be the number one training data supplier for artificial intelligence initiatives.
