Mariia Rodina
Angestellt, Machine Learning Engineer, Epoch8
Yerevan, Armenien
Über mich
Innovative and results-driven Machine Learning Engineer with 5 years experience, skilled in Python, AQL, SQL and adept in cutting-edge ML and AI frameworks. Proven track record of mentoring, integrating AI solutions into various services and executing innovative research sprints. Strong background in deploying AI-assisted tools and NLP techniques to significantly elevate performance. I am confident in my ability to bring valuable expertise and innovation to a new role, and eager for the relocation opportunity.
Werdegang
Berufserfahrung von Mariia Rodina
Bis heute 1 Jahr und 6 Monate, seit Jan. 2023
Machine Learning Engineer
Epoch8
* Designed and applied an external FAQ service that ensured the delivery of new data to the chatbot within 15 minutes, with the top three bot responses achieving over 90% accuracy during validation; * Leveraged a text communication channel as a data source to develop an AI-based text assistant for searching a formalized data source with a response speed of 2-5 ms; * Mentored junior team members, providing guidance on best practices, coding techniques, and machine learning model development;
9 Monate, Mai 2022 - Jan. 2023
Machine Learning Engineer
Smart Consulting
* Pioneered the automation of graph database population and semantic search, dramatically reducing the amount of man hours required for data labelling. This approach resulted in several-fold savings in time and resources, significantly boosting overall productivity. * Utilized cutting-edge techniques in NLP, such as metric learning on texts, topic modeling, and entity extraction, employing pre-trained, fine-tuned, or trained from scratch models.
3 Jahre, Juni 2019 - Mai 2022
Machine Learning Engineer
CJSC ”NeoCor”
* Designed and developed the warehouse demand forecaster, that reduced redundant warehouse processes by 1.8 times. * Designed and developed the multitasked service supporting a sales department with the zero-shot contract predictor and the tender’s proposal system (partial implementations) in it. * Collected a dataset with a million+ of records from lightly structured sources (web scraping, Selenium + BeautifulSoup libs). Labeled data with an active learning algorithm.
Sprachen
Englisch
Fließend
Russisch
Muttersprache