NeuralSpace
How a language barrier while traveling led Ayushman and Felix to form NeuralSpace.
What is your background?
Ayushman and Felix, originally from India and Germany, respectively, are both engineers with a passion for exploring how cultures and languages have emerged together around the world. They love reading about foreign cultures and exploring them in person by traveling to many different countries.
Professionally, Ayushman was first active in Germany after his Master in Computer Science at the Technical University of Kaiserslautern (where he was connected to Felix through Kumar Shridhar, NeuralSpace’s third founder). He then worked for the German Research Institute for Artificial Intelligence and Insiders Technologies before he stepped out of Europe to work for the NLP/language technology group at Montreal-based Element AI, which was co-founded by one of the fathers of deep learning, Yoshua Bengio. In November 2019, he took the bold step and quit at Element AI to work on NeuralSpace that he co-founded with Kumar and Felix a few months earlier.
Felix finished his Masters in Engineering at the Technical University of Denmark in 2018 before he moved to the UK to pursue a Ph.D. in Statistics at Imperial College London. In his research group, he collaborates with scientists who have studied and worked at MIT, Stanford, and Caltech.
What problem are you solving? Did you face it personally?
Ayushman and Felix (in fact also Kumar) have already traveled to many countries where they faced enormous difficulties due to not speaking the local language. Buying a train ticket, a seamless task for anyone traveling to an English-speaking country becomes very difficult suddenly. Even the technology counterpart of any ticket agent, an apparently very smart chatbot, can most of the time not understand what we want when we write in a different language, let alone voice assistants.
In summary, current NLP solutions majorly focus on one of the few high-resource languages like English, Spanish or German although there are about 3 billion local (aka low-resource) language speakers (mainly in Asia and Africa) globally. Such a large portion of the world population is still underserved by NLP systems because of various challenges that developers face when building NLP systems for low-resource languages around data quality.
Here is where NeuralSpace solves the problem by enabling access to any software or mobile developer to deploy NLP solutions in low-resource languages. It helps to break down the language barrier and bridge this huge gap by enabling people to access the Internet in the language and mode of their choice (voice or text). Through the NeuralSpace Platform, any software product can reach over 6 billion people in their native language.
What is your solution?
NeuralSpace is a Software as a Service (SaaS) platform which offers developers a suite of APIs for NLP tasks that you can use without having any machine learning or data science knowledge. The NeuralSpace Platform currently consists of 5 Applications: NeuraLingo, Neural Machine Translation, Transliteration, NeuralAug, and Language Detection. NeuralSpace offers language support of 80+ languages spoken across India, South East Asia, Africa, Middle East, Scandinavia, and Eastern Europe, aka, low-resource languages.
Why do you believe that this is the perfect time to build this product?
Even today, there are billions for whom accessing the Internet is still a distant dream, not because of a lack of connection, but because there are close to no solutions available in a language they can comprehend. This goes from mobile apps that allow you to order items online which people cannot buy in their local villages, to booking an appointment for a covid-19 vaccine through an app. Given our more and more digital world, access to the Internet needs to be given to every citizen regardless of the language they speak.
Why are you excited about the industry you are building in?
In developing countries, where most of the low-resource languages are spoken, the NLP industry has classically been a B2B market where custom software is built for each customer. It is not the case that customers do not employ software or mobile developers, they just do not have the skillset that is required to build advanced machine learning and NLP systems. The NeuralSpace Platform takes out all of the complexities that are associated with building an NLP system, like a voice command feature, into software or mobile application. The in-house development of AutoNLP and AutoMLOps makes it possible for developers to focus on the integration of the NLP system into their applications.
What has been the biggest lesson you have learned so far on your journey?
A founding team that is humbly willing to learn never portrays itself as being flawless, and respects each other greatly can overcome almost any challenge they face.
Who have been your biggest supporters up-today?
David Rangel at Merus Capital, Vijay Tirathrai at Techstars, Sami Abou Saab, and many more.
What do you need help with?
We are always eager to learn and if anyone has material to study on what we need to consider in building a SaaS product, please let us know!