In my PhD research I applied deep learning models for information extraction from conversations. I used pre-trained transformers for creating contextualized representation of words and then I was adding features, specific to conversations (speaker information and dialogue structure). One of my models was based on distantly supervised BERT fine-tuning using reinforcement learning feedback to predict long-tailed personal attributes in a zero-shot setup.
exciting research challenges, work in fast-paced environment, taking the ownership and responsibility in the project. Making impacts into real product and influence real peoples' lives. Work with various teams in cross-functional way.
- United States
- United Kingdom
Expected Base Salary
Total Professional Experience
Send a connection request to the candidate to get their contact details.Contact Candidate