AB Team

Data Scientist

- Development of a pipeline based on GRU to determine the position of the forearm in
space based on the time series of the positions of the biceps and hand. Now the model is
used in vr-helmets of one of the manufacturers
- Development of a software product for analyzing the states of an X-ray machine
operator in public places. The product is being tested at all airports in Moscow and other
cities.
- Development of a high-precision input-invariant DL-pipeline that finds containers,
bladed weapons and firearms on x-rays of luggage. The product is being tested at
Vnukovo and Domodedovo airports and used at Pulkovo and the Hermitage.
- Development of a product for recognition of printed and handwritten documents with
different structures. The product is actively used by many clients: Alfa Bank,
Renaissance, Rosseti, Skyeng, Yandex and others. He was engaged in improving the
quality of the product on the main types of documents: invoices, 2NDFL. I also
simultaneously developed a branch of the product for recognition of US invoices: under
my leadership, 2-3 people worked on the Symbol Detector + OCR + NER pipeline.
- Development of an embedded module for recognizing a person by a picture of his face.
Now the module is built into customer services: Social Links, Coral Club.
- Development of an MVP product for face swap detection in photos and videos
generated using DeepFakes and FaceSwap technologies.

Main programming languages: Python