DeepPavlov Internship Program

Are you obsessed with the possibilities of the emerging fields of artificial intelligence? Do you have ideas on how to improve dialog systems for everyone? Are you ready to make an impact across the world? Great, then join us!

We’re looking for the aspiring interns in NLP who have a talent for problem solving, are passionate about Conversational AI technology, and offer diverse perspectives.

Who are we?

At DeepPavlov, our focus is on building a Conversational AI technology stack towards solving the Holy Grail of NLP: computers should achieve human-like comprehension of texts/languages. To achieve this truly hard task, we invest to spread across 3 parts of the technology stack, including Library, Dream, and Agent.

We believe that artificial intelligence technologies such as natural language processing will have a massive impact on the way we live, and we want to make sure that we create a positive impact on the world with these AI services.

DeepPavlov is famous for its organization of ConvAI and ConvAI 2 challenges, regularly participates in academic (EMNLP, ACL, etc.) and business conferences (NVIDIA GTC, ODSC West etc.) across the world, and is proud of its participation in Amazon Alexa Prize Socialbot Grand Challenges.

Who are you?

You’re currently studying Computer Science, Artificial Intelligence or a related field, either a bachelor or master degree, and are looking for an internship to get some key research experience, and/or to do your bachelor project or master thesis.

You are interested in exploring the state-of-the-art technology in machine learning, and applying it to develop open-source projects. Together with the team we are working on projects that use natural language processing.

You’ll join a small but fast-growing data science team, responsible for implementation of various AI projects. Some examples of tasks would be reviewing literature, setting up experiments, contributing to the code base, and collaborating closely with other teams to make things happen.

… and some more points:

• experience with modern production DS stack: Python 3, pandas, numpy, PyTorch, Keras/TF
• understanding of computer science, theory of statistics & probability, ML & DL

note: If you don’t feel that you perfectly fit the description above, but you have some other background or skills that you think could be of interest, please still get in touch!

Which projects could you work on?

We have several different projects, which are active during Summer and Fall 2021:

• Build A Digital Twin of Historical Character
  This project aims at creating a (limited) digital twin of a known historical character. This twin must have some kind of a knowledge base, to speak like an original.
• Implement New Loss Function
  This project is focused on adding minor changes to the training texts, and penalizing the model for sensitivity to them. It will be necessary to see how the losses are implemented in modern neural networks and write our own loss-regularization, looking not only at the output of the model, but also at its intermediate layers.
• Extract Knowledge Graph From Language Model
  This project aims at reviewing language models to evaluate the relationship of input words with each other. Check the possibility to get out of this a graph similar to ConceptNet or WikiData.
• DeepPavlov Integration with Weights and Biases
  This project is focused on adapting our open-source DeepPavlov Library for running on platform — Weights and Biases to visualize ML experiments.
• Discourse Management Recommendation System for Socialbots
  This project helps us to build a bridge between linguistics and NLP with the focus on advanced utterance classification.
• Russian MultiWOz
  This project aims at developing a new model based on our data to submit as Russian version of the MultiWOZ benchmark.
• Adapting Deepy For Embedded Devices
  This project is focused on adapting our open-source Multiskill AI Assistant, Deepy, for running on embedded devices like NVIDIA Jetson Xavier NX.
• Migration of TF 1.0 Models to PyTorch
  This project aims at driving the development and release of the v1.0 of the DeepPavlov Library by migrating selected ML models to PyTorch.

note: If you don’t feel that you perfectly fit the description above, but you have some other ideas that you think could be of interest, please still get in touch!

What do you get?

As an intern, you will work on some of the most ambitious technical problems, develop new ML solutions that will impact future DeepPavlov products stack. All interns are paired with a mentor and will participate directly in DeepPavlov’s groundbreaking work. There are no restrictions on publications based on internships.

The duration of this internship is 2–3 months with Summer, Fall, and Winter start dates.

Given the present COVID-19 situation, you’ll start working remotely. We are having regular meetings, discussions and online fun activities, and are hoping to be back in the office once the situation is again under control.

Conclusion

If this all sounds interesting to you, then please fill the form at our official website. We would love to hear from you 🙂

You can read more about us in our official blog. Also, feel free to test DeepPavlov by using our demo. And don’t forget DeepPavlov has a dedicated forum, where any questions are welcome.