Machine learning Teacher

• Mar 2023 - Present

I work with undergraduates of the Machine Learning direction. I work with undergraduates of the Machine Learning direction.

Technical Lead (Data Science)

• Jul 2021 - Present

I lead a team of 12 developers who are engaged in the full cycle of development, implementation, and support of high-load services based on machine learning and search. Main directions: Recommendation systems for B2C and B2B users. Salary prediction based on resume\vacancies Prediction of skills Automatic recruitment systems Extracting NER from unstructured text. Personalized ranking of search results We work on augmented CRISP-DM from the point of view of machine learning. In terms of SCRAM and OCR software development processes. Technology Stack: Python 3.7+, Go, Kafka, Faust, FastAPI, Hadoop, Spark, Airflow, PostgreSQL, ClickHouse, Huggingface Transformers|Tokenizer|Datasets, Tensorflow, PyTorch, aiohttp, pydantic, FAISS, Neo4j, Elasticsearch, DVC, Kubernetes, GitlabCI, etc. The team has Open Sorce projects. We have assembled a 2TB dataset specific to the HR domain. And we trained a multilingual BERT-base model - HRBert, using the power of the Christofari Neo supercomputer. Its distilled version is laid out in Huggingface HUB https://huggingface.co/RabotaRu/HRBert-mini We participate as a track holder in the AIIJC Machine Learning Competition (https://youtu.be/Gz_C8jD87Lk?list=LL&t=9660 ) And we perform at meetups: (https://www.youtube.com/watch?v=AOU9YBgUVkE&list=LL&index=51&t=127s https://ods.ai/events/spb_meetup_2022-04-29 ) My speech at HighLoad++ is about how we make recommendation services in "Rabota.ru" https://www.youtube.com/watch?v=O589t08FfIY&feature=youtu.be Show less

Author of courses on Machine Learning

• Apr 2021 - Dec 2022

Development of training modules on machine learning and data analysis for advanced level Development of training modules on machine learning and data analysis for advanced level

Senior ML engineer

• Mar 2020 - Nov 2021

My responsibilities as a Senior ML engineer include: - Researching advanced approaches in machine learning; - Identifying opportunities for customers in the field of machine learning; - Help in the development of the solution architecture; - Deep data analysis; - Full cycle of ML services development, including data engineering, deployment and automation pipelines, solution analysis, testing solutions from objective metrics to custom business-specific metrics and A\B tests, IaC development, etc. At the moment, it was possible to successfully complete and run in the prod: - Financial consulting project of personalized investment recommendations. - Advanced data deduplication service for the largest asset management ecosystem for the commercial vehicle industry - Several recommendation system projects from the flower sales network to the largest cable operator in the United States. - Creating dashboards for a university in the southern United States. - Projects in the field of the Internet of Things. - Project in the field of machine learning in medicine - Project for selecting resumes for selected vacancies. For more information about completed projects, see the Projects section at the bottom of my profile Core technologies: AWS forecast, AWS Batch, Lake formation, Amazon Athena AWS Sagemaker, AWS Step Functions, AWS GLUE, AWS ECR, AWS Lambda, AWS API Gateway, AWS Personalize, Amazon Transcribe, AWS Neptune, Amazon Comprehend Medical, AWS IoT Core, AWS Textract, IoT Analytics, IoT Device Management, IoT Events, Amazon QuickSight, AWS Redshift, AWS Fsx for luster, AWS Pinpoint, AWS SNS, AWS CloudWatch, AWS Kinesis Firehouse, huggingface transformers, RNNnoise, SentenceTransformers, Textrank, Nltk and others. Programming languages used: Python, Swift , SQL, Serverless, Gremlin, Node.js Show less

Principal ml engineer

• Oct 2020 - Sep 2021

Development of a cloud service architecture based on machine learning to determine skin type. Core technologies: AWS Sagemaker, AWS Step Functions, AWS ECR, AWS Lambda, AWS API Gateway and others. Programming languages used: Python, SQL, Serverless Conducting research on voice cloning methods. Creating a prototype. Technologies: AWS Sagemaker Programming languages used: Python, C Development of a cloud service architecture based on machine learning to determine skin type. Core technologies: AWS Sagemaker, AWS Step Functions, AWS ECR, AWS Lambda, AWS API Gateway and others. Programming languages used: Python, SQL, Serverless Conducting research on voice cloning methods. Creating a prototype. Technologies: AWS Sagemaker Programming languages used: Python, C

ML engineer (NLP)

• Sep 2018 - May 2020

- Development and implementation of text document classi cation system based on FastText embedding + bidirectional neural network on LSTM cells - Development and implementation of a system for Named Entity Recognition in text documents. Based on a combination of Deep learning (BiLSTM) and rules system (Tomita-Parser) - Software development for unstructured data parsing - Development of a passport verification service based on convolutional neural networks. - Development of a service for checking voice calls, with automatic search for negative subtexts in the conversation. Core technologies: flask, tensorflow, yargy, NLTK, pymorphy2, pykaldi, pandas and others Programming languages: Python, SQL Show less

Data Scientist (Scoring)

• Feb 2019 - Mar 2020

Completed projects: - Development of credit scoring models, using credit Bureau data, website and information from social networks - Development of customer segmentation models, based on their pro les and analysis of their behavior - Research trends and technologies in the eld of Data Science, AI and Machine Learning for their subsequent application Core technologies: LightGBM, Flask, SHAP, Lime, Tensorflow, Posgresql, pandas, numpy, scipy, scikit-learn, Docker and others Programming languages: Python, SQL Show less

Electronics engineer/ML engineer

• Sep 2017 - Aug 2018

- Developed and implemented the control system technospheric States of the pipe based on the CNN - Developed and implemented a system of input control of documents, based on CNN - Modeling of operating modes and development of Biterm electric coupling welding machine Core technologies: Tensorflow(Keras), pandas, scikit-learn Programming languages: Python, C - Developed and implemented the control system technospheric States of the pipe based on the CNN - Developed and implemented a system of input control of documents, based on CNN - Modeling of operating modes and development of Biterm electric coupling welding machine Core technologies: Tensorflow(Keras), pandas, scikit-learn Programming languages: Python, C

Research assistant

• Sep 2008 - Aug 2016

- Development and implementation of voice control system for industrial equipment based on hidden Markov models - Implementation of digital filtering algorithms, and their adaptation to work with digital signal processors Core Technologies: SciPy, scikit-learn, DSP(Matlab, Simulink) Programming languages: Python, Matlab, C++, Assembler(PIC microcontroller) - Development and implementation of voice control system for industrial equipment based on hidden Markov models - Implementation of digital filtering algorithms, and their adaptation to work with digital signal processors Core Technologies: SciPy, scikit-learn, DSP(Matlab, Simulink) Programming languages: Python, Matlab, C++, Assembler(PIC microcontroller)

Software sales Manager

• Aug 2011 - Sep 2012

Specialist of sales Department

• Apr 2011 - Aug 2011