- Ważna jeszcze 23 dnido: 12 lis 2020
- Praca zdalna
- B2B contract
- specialist (Mid / Regular), senior specialist (Senior)
Our conversational AI visual programming language. Together with the rest of the team you will be responsible for designing the language and its runtime behavior. We need to allow human Dialog Owners to manage the complexity as well as deliver unequivocally the best solution for building machine-to-human dialogs. We are currently using a state-machine approach but at some point AI will be involved in executing dialogs as well as generating language.
Mini-ASR development - ASRs are optimized for dictation. In conversations there are many phrases that very often get wrongly transcripted by generic third-party systems. That's why we develop custom mini-ASRs which provide proper transcription for NLP models based on our production data. Challenges: Multi-lang and constant improvement of dataset size vs. accuracy ratio.
Simple and advanced NLP models - models including NER, Regexp+deep learning, in general the most important matter in terms of AI is fastest possible generalization. Challenges: Multi- lang and improving generalization by using multiple approaches incl. transfer learning, multi-task learning, and others.
End-of-speech - allows to maintain natural feeling of conversation through minimizing time of voicebot reaction. Challenges: Multi-lang, improving confidence through working on neural network and data: audio, text, tabular data
Voicemail detection - important for proper qualification of outbound calls. Challenges: Proper classification of new undetected cases constantly showing up in production.
Software 2.0 - In order to make conversations more personal we need to master the process of continually improving our systems according to data from production. Challenges: currently there are too many manual steps in our process. There’s a huge potential to use customer’s resources - they want control and Quality Assurance process could integrate data tagging function (yes we know that is risky - we have to control it).
You have a proof of technical excellence in the field of Artificial Intelligence.
You have strong interest in NLP.
You are self-driven. Autonomous.
You like to decompose problems and understand them through first principles
You have implemented novel neural network architectures based on white-papers.
You are proficient in designing and creating custom neural networks architectures.
You have experience developing NLP models.
You possess skills required to practice MLOps such as deep understanding of infrastructure .
You have good grasp of data processing pipeline design.
You have knowledge and intuition to design data tagging processes.
You have strong Python programming skills (familiarity with C++ or functional programming is a plus).
You have good understanding of computer science fundamentals, especially around performance.
What we offer
Working in a spin-off project within a company with stable market positition
Challenging tasks in a rally to develop Conversational AI which will be a preferred to human agents by both companies and their users
Significant impact on defining our product and the category of digital speech interface
Developing commercially proven AI technology across Enterprise Customers (currently 60 NLP models and mini-ASRs used in production every day)
Working on our own infrastructure (currently 8GPU cards)
Being one of the first members of the team - you will co-create our culture
Working with some of the smartest and the most talented people, who are synchronized but no one tells them what to do
Flexible vacation policy
Dialo creates AI-powered virtual agents and tools around them to enable business customers scaling personalized conversations which help them build great relationships with their users. Currently our system processes 2000h of conversations a month and is fully powered by our own stack.