Make AI accesible!
Greetings! My name is Igor Safronov, I am a Python programmer. I specialize in the development of artificial intelligence in the field of natural languages (NLP) and time series forecasting.
I understand the principle of operation of a neural network, the logic of building its architecture and the influence of the selected variables on the final result. I studied at Harvard with a degree in CS50 Artificial intelligence and am constantly expanding my knowledge in this area by watching MIT Deep learning lectures and reading articles on Medium.com.
Here is a list of tasks which i handle with:
- fine tuning of pretrained language models (BERT, BART, DeBERTa, RoBERTa, GPT3, MT5, T5, XLNet) for various tasks:
• text classification (sentimental analysis, categorization)
• zero-shot categorization
• named entities and relationships recognition (NER, NEL)
• text generation
• summarisation
• question answering (QA, table QA)
• texy similarity
- integration of AI with chat-bots;
- manual markup of datasets for AI training;
- preliminary processing of texts (cleaning, tokenization, regular expressions);
- development and training of convolutional and recurrent neural networks from scratch for tasks:
• data classification (category prediction based on passed parametres)
• regression (prediction of numerical values)
My stack: Tensorflow, Transformers, Pandas, Numpy, scikit-learn, scipy, nltk, Matplotlib, Gensim, label studio, hugginface, github
Work Terms
Crypto currency wallets/ Binance p2p. Payment upon completion.
Attachments (Click to Preview)
-
-