1. Trained a model to estimate the probability of conversion of insurance applications which allowed to increase conversion rate by 2% | python, docker, fastapi, catboost, sql | | 1 PM, 1 ML Engineer, 1 full-stack
2. Created functionality for matching e-learning content with the categories in which it should be placed with an accuracy of about 90% | python, docker, fastapi, catboost, transformers | 1 PM, 2 ML Engineer, 2 full-stack
3. Built a system for detecting logos and matching them with various brands to estimate advertising screen time with an accuracy of 83% | python, docker, flask, tensorflow, pytorch | | 1 PM, 1 ML Engineer, 2 full-stack
4. Built a RAG chatbot to answer the questions from doctor/patient conversation transcripts | python, docker, flask, langchain, pytorch, whisper | 1 PM, 1 ML Engineer, 1 full-stack
5. Created a prototype of a system for parsing large financial reports to extract from them the values ​ of certain metrics with an accuracy of 79% | python, openai | 1 ML Engineer
6. Created a prototype system for estimating the protective coverage area of ​ medical flasks with per pixel accuracy of 87% | python, docker, tensorflow | 1 ML Engineer
7. Created a prototype of a system for analyzing the contents of construction plans (finding and counting elements according
to the references) | python, transformers | 1 ML Engineer