Experience

• 

  Paralles - mipt center

  Jan 2006 - Jun 2010

  Researcher

  project “Virtualization Technology in GRID”

• 

  Acronis

  Jan 2006 - Jan 2007

  Engineer
• 

  Space research institute of the russian academy of sciences (iki ran)

  Jan 2009 - now

  Data analysis from X-ray observatories (INTEGRAL, RXTE, SWIFT). Fourier analysis, fractal analysis, spectral analysis (Heasoft, OSA, etc).Programming: Python (NumPy, SciPy), Matlab, R.List of papers:Grebenev, S.A., Prosvetov, A.V., Sunyaev, R.A. “Second outburst of the black-hole candidate MAXI J1836-194 from SWIFT and INTEGRAL observations”, Pis'ma v Astron. Zhurn., 2013, 39, 415-422 (Astron. Lett., 2013, 39, 367-374)Grebenev, S.A., Prosvetov, A.V., Burenin, R.A. "Broad-band emission spectrum of the X-ray nova SWIFT J174510.8-262411 at the decaying stage of its outburst'', Pis'ma v Astron. Zhurn., 2014, 40, 198-204 (Astron. Lett., 2014, 40, 171-176)Prosvetov, A.V., Grebenev, S.A. "Fractal analysis of light-curves from the microquasar GX339-4", Pis'ma v Astron. Zhurn., 2014, 40, 761-773 (Astron. Lett., 2014, 40, 691-703)Prosvetov, A.V., Grebenev, S.A. "QPO and LFN in the power spectrum of fast variability of the X-ray nova SWIFT J174510.8-262411'', Pis'ma v Astron. Zhurn., 2015, 41, 1-14)Prosvetov, A.V., Grebenev, S.A. "QPO and VLFN in the power spectrum of V404 Cyg detected with INTEGRAL", The Astronomer's Telegram, #7726Prosvetov, A.V., Grebenev, S.A. "LFN, QPO and fractal dimension of light-curves from X-ray binaries", Proceedings of International Conference MSS-14 Свернуть

  • Leading Mathematician

    Apr 2017 - now

  • Researcher

    Jan 2009 - Apr 2017
• 

  Moscow institute of physics and technology (state university) - mipt, phystech

  Jun 2009 - Dec 2012

  Ph.d. candidate
• 

  Runa capital

  Aug 2010 - Oct 2011

  Analyst
• 

  Clever data

  Mar 2016 - Jun 2021

  Leading the group of Data Scientists in the following areas:- Recommendation Systems;- Preparing system of automatic marketing;- Text classification, Topic Modelling;- Audience segmentation;- Predicting the behavior of site visitors using neural networks;- Graph analysis; - Performing experiments with Neural Networks.Conducting lectures and training in the Data Science in National Research University Higher School of Economics .Participation in conferences. Won Best Oral Presenters Award in international conference FSDM 2019 (Japan).List of recent Papers:GAN for Recommendation System, A. V. Prosvetov 2019 J. Phys.: Conf. Ser. 1405 012005Fractal techniques as alternative of Fourier transform for fast search of periodic events in time series, A Prosvetov, Journal of Physics: Conference Series, Volume 913, conference 1The comparison of autoencoder architectures in improving of prediction models, A. V. Prosvetov,Journal of Physics: Conference Series, Volume 1117, conference 1 Свернуть Data analysis:- Recommendation system;- Segments building and performing custom audience analysis;- Predictive modeling (LTV, churn, anti-fraud, Loyalty, best communication channel and etc);- Natural Language Processing (topic modeling, text classification, fuzzy string matching and etc);- Analysis of blogs and development of tools for recommending a better blogger for product promotion based on product information; - Lookalike modeling;- Deep Learning models for prediction of time series (Tensorflow, Keras, etc);Conducting lectures and training in the Data Science and Big Data (Netology). Свернуть

  • Lead Data Scientist

    Feb 2018 - Jun 2021

  • Data Scientist

    Mar 2016 - Feb 2018
• 

  Ланит

  Jul 2019 - Jun 2021

  Chief data scientist

  Leading the group of Data Scientists in the following areas:- Predictive maintenance for complex technical systems, including anomaly detection and long-term degradation forecasting.- Time series analysis of diverse multivariate time series describing the behavior of complex systems.- Natural Language Processing for various tasks using machine learning in technical texts.

• 

  Yandex.taxi

  Jul 2021 - Oct 2022

  Lead ml developer

  - Conducted behavior estimation and prediction to improve decision-making processes and enhance customer insights.- Developed advanced behavior embedding techniques to effectively capture and analyze user behavior patterns.- Implemented uplift modeling strategies to optimize marketing efforts and increase customer engagement.- Predicted B2B bank scoring outcomes using data-driven models to assess creditworthiness and financial risk.- Engaged users through financial incentives calculated via machine learning to drive participation and retention.https://highload.ru/foundation/2022/abstracts/9009 Свернуть

• 

  Yandex

  Oct 2022 - now

  Lead ml develper

  - Executed 3D scene reconstruction in sports events using machine learning techniques.- Developed a statistical analysis tool for sports analytics to enhance data-driven insights.- Created methods for assessing the accuracy of collected sports statistics.- Developed a method for 3D ball localization in soccer matches captured by a single camera.