Nullius in Verba (NIV) is a software company based in Tirana, Albania. We are a team of skilled professionals who are committed to delivering innovative software solutions to our clients. Our skills allow us to leverage the development of cutting-edge software solutions that meet the evolving client’s needs.

We pride ourselves on our expertise in a range of technologies and programming languages and are constantly staying up-to-date with the latest industry trends and best practices to ensure that we are delivering the best solutions. Whether you need a web or app application or design system, we have the skills and experience to deliver high-quality results.

Services we offer:

Web Development

Whether it is a simple website or a complex one, we can get the work done. By using Worpress or Laravel we can create websites according to the customer needs

App Development

In case of a need of an application, we will most likely use Flutter and develop it in record time without having the need of spending too much on it.

Design

Branding / Art Direction / Logo Design / Packaging Design / Video Editing / Motion / Social Media

Machine Learning

For machine learning we use Pytorch for training, fine - tuning, validating and inference on your machine learning methods and AWS Sagemaker which will make your model public.

We have also created a online marketplace platform which is recently launched in the Albanian market, where anyone can post in different attributes whether it is Housing, Cars, Jobs, Market, Services and Community. The platform name is migjen.ai .

But not only that, we are creating (TTS, STT, and LLM), models which can transcribe audio, perform text speech and perform actions (Large Language Model). We are in the process of fine-tuning the model which can perform the task in Albanian, constructing the dataset of 3’000 hours of transcribed recordings. The models use the same architecture as Whisper, and we have in possession models of 15B parameters. The current model can perform in 96 different languages. We are in the process of fine-tuning the model which can perform the task in Albanian, constructing the dataset of 3’000 hours of transcribed recordings. We possess the models (TTS, STT, and LLM), and know-how, and have demos we can showcase.

The link below is early stage of speech to text, the demo is uploaded in hugging face so you can access it:

https://huggingface.co/spaces/niv-al/wsq

For anything else contact us.

Nullius in Verba

Write a review for Nullius in Verba

Please fill the form below

Budget

Quality

Schedule

Communication