Expert Details

Expert in Machine Learning, Applied Statistics, Particle Physics, Python, C++

Expert ID: 736571 Illinois, USA

Request Expert

Machine learning, Applied Statistics, Particle Physics, Python, C++.

Big experience in experimental research, data analysis and modeling. Co-author of many software packages in physics experiments, industry and SaaS analytical platform developed in the digital advertising eld. Supervision of undergraduate, Ph.D. students, data scientists and interns in various projects. Primary author of many papers published in major peer-reviewed physics journals with application of machine learning algorithms in physics experiments. He also gave a few
talks at international conferences and workshops.


Areas of Expertise
 Experimental Data Analysis  Statistical Analysis  Collaborative Work
 Particle Physics  Programming  Project Leadership
 Algorithms Development  Machine Learning  Teaching & Supervision
Skills
{ Programming in Python, R, C++, Scala, Java, Fortran;
{ MatLab, ROOT (data processing and analysis framework), PAW (physics analysis workstation), RStudio;
{ SQL (incl. PostgreSQL, MySQL, SnowSQL), noSQL (Mongo, Redis, HBase, Cassandra, Neo4j) data bases;
{ Big data processing using Hadoop, HDFS, programming using Spark, Hive, MapReduce, Pig;
{ Machine learning using scikit-learn, MLLib, TensorFlow, Keras, PyTorch, H2O, distributed deep learning;
{ Natural language processing [incl. Gensim/NLTK/SpaCy/ULMFiT/ELMo/TextRank; GloVe/Word2Vec];
{ Computer vision [incl. OpenCV, YOLO/ResNet/SSD, OCR];
{ AWS Cloud (RDS, Amazon S3, EC2&ECR, Elastic Beanstalk, Lambda, SageMaker);
{ Google Cloud (Compute, Kubernetes and Cloud ML engines, BigQuery, DataLab);
{ Azure Cloud (Databricks, Delta Lake, Azure ML, Synapse Analytics);
{ Data visualization (Tableau, PowerBI, Python&R libraries, e.g. Plotly, Dash, Shiny);
{ Live data monitoring, writing software codes to control hardware operation.

Professional Achievements
Research, Development in Scienti c Laboratory Environment
{ Work at Fermi National Accelerator Laboratory (Fermilab, USA), 2000{2014; European Organization for
Nuclear Research (CERN, Switzerland), 1997{2011; Deutsches Elektronen-Synchrotron (Germany), 1998{1999.
Project Supervision, Management and Task Leadership
{ Leader of data science group: Strike Social, 2016{2018;
{ Leader of two research physics groups: Fermilab, 2009{2014; two algorithm groups: Fermilab, 2006{2010;

Expert may consult nationally and internationally, and is also local to the following cities: AURORA, IL - Naperville, IL - Chicago, IL

Request Expert

Education

Year Degree Subject Institution
Year: 2004 Degree: Ph.D. Subject: Elementary Particle Physics Institution: JINR
Year: 1993 Degree: M.S. Subject: Theoretical and Nuclear Physics Institution: Saratov State University

Work History

Years Employer Title Department Responsibilities
Years: 2019 to Present Employer: Undisclosed Title: Sr. Lead Data Scientist Department: Responsibilities: Optimization of manufacturing commercial trucks and buses. Support of safe driving, processing of IoT, telematics data, prediction of repairs and maintenance, anomaly detection, estimate of vehicle remaining useful time. Supply chain optimization. Big Data processing using CDSW, Hive, Impala, and streaming data in Azure Cloud.
Years: 2019 to Present Employer: Undisclosed Title: Head of Data Sceince/ Advisor Department: Responsibilities: Head of a team specializing in following directions: predictive maintenance and repairs, supply chain optimization, advertisement, eCommerce, document analytics and transcriptions of medical audio.
Years: 2016 to 2019 Employer: Strike Social Title: Sr. Data Scientist Department: Responsibilities: - Development of recommendation and optimization tools for video advertising campaigns (with multimillion dollar budgets) at the largest YouTube advertising and data-driven company.
- Created online AI tools that support video advertising campaigns on DoubleClick, Adwords and Facebook platforms with multiple KPIs in a real time.
- Development and support of a number of online predictive models: How data scientists can improve social media advertising
- The algorithms work in autonomous manner as a part of a new SaaS analytical platform with his primary contribution, and allowed to automate and significantly boost efficiency of advertising campaigns.
- Development of a few NLP-based technics to identify good and bad YouTube channels, and channel recommendation system.
- Developed algorithms that select 99% brand-safe channels according to IAS standards
- Research of object identification on videos.
- Customer analytics, LTV modeling and predictions, optimization of promotions.
Years: 2014 to 2016 Employer: Centro Title: Sr. Data Scientist Department: Responsibilities: - Development of recommendation system for a programmable media platform, optimization of advertising campaigns, programming, statistical data analysis.
- Development of audience, clicks and viewability prediction models.
- Analysis of user cookie-level data and real-time bidding data (multi-TB scale) using Amazon S3, Hadoop cluster with Spark, Spark MLLib, and Hive. Social media analytics. Supervision of two interns (both found Data Scientist positions same year).
Years: 2013 to 2014 Employer: University of Virginia / Fermilab Title: Sr. Research Scientist Department: Responsibilities: Head of Quantum Chromodynamics and Jet Energy Scale physics groups.
Publication and presentation of the group results at workshops and conferences.
Years: 2010 to 2013 Employer: Florida State University / Fermilab Title: Research Scientist Department: Responsibilities: Head of physics group “Quantum Chromodynamics” (August 2009 – October 2014).
The group (about 70-80 scientists) studies processes with parton strong interactions. During his leadership of the group, 25 measurements have been published and submitted to physics journals. As a leader of the group, was responsible for selecting physics topics, communication with analysers, their technical and other support, organizing and running the group meetings, preparation of physics papers, presentation of results at the collaboration workshops.

Convener of physics group “Jet Energy Scale” (May 2011 – October 2014).
The goal of the group (with about 30 scientists) was developing methods and providing energy scale corrections for hadronic jets identified in the D0 calorimeter. The accuracy of the corrections is crucial for most of the experimental results. The jet energy scale method developed in D0 is the most precise in the world (with 1 − 2% accuracy in the most of phase space).
Results of developments during a few years have been published.
Years: 2006 to 2010 Employer: Kansas State University / Fermilab Title: Research Scientist Department: Responsibilities: Head of Electron and Photon Identification group (January 2007 – August 2009).
The main goal of the group (with about 25 members) is providing tools and methods for a reliable identification of electromagnetic objects (electrons and photons) in the D0 detector. As a leader of the group, was responsible for developing and improving electron and photon identification criteria, their software implementation and support, and periodic validation of the identification efficiencies. Introduction of multivariate techniques allowed to decrease background rates up to 40%. All algorithms, criteria and results of their application was published.
Optimization of D0 electromagnetic triggers.
Calibration of CMS EM and hadronic calorimeters using test beam data.
Development of Level 1 and High Level Triggers, identification tools for low energy electrons.
Development of algorithm for reconstruction of clusters in the CMS preshower detector (implemented in the official reconstruction code).
Development of a method for improving energy resolution of electrons/photons in the CMS calorimeter (implemented in the official CMS software).
Co-author of a few official CMS software packages.
Years: 2000 to 1964 Employer: Joint Institute for Nuclear Research Title: Scientific Associate Department: Laboratory of Nuclear Problems Responsibilities: Preparation (Monte Carlo modeling, analysis of beam test data) of CMS experiment.
Physics analyses of D0 data at the Tevatron.
Work at D0 Muon detector group (2001 – 2004): support of software tools for the muon system monitoring, work in EMID and QCD groups.
Teaching at Dubna University and JINR University Center.
Years: 1995 to 2000 Employer: Joint Institute for Nuclear Research Title: Scientific Associate Department: Laboratory of High Energy Physics Responsibilities: Preparation of CMS experiment: detector modeling with GEANT package, test beam data analyses for hadron calorimeter, development of a method for jet energy calibration and measuring gluon distribution in a proton.
Work at HERA-B Experiment (1998-1999, Deutsches Elektronen-Synchrotron, Germany).
Work in the transition radiation detector and EM calorimeter groups. Detector simulation of the physics processes using GEANT; development of tools for a global alignment of the HERA-B subdetectors.

Career Accomplishments

Licenses / Certifications
- Complete Tableau Bootcamp for Data Visualisation, Oct 2019. No Expiration Date
- Complete Guide to Tensorflow for Deep Learning with Python
- Deep Learning: Advanced Computer Vision
- Artificial Intelligence: Reinforcement Learning in Python, Feb 2018. No Expiration Date
- Data Science, Issuing authority University of Washington, Jul 2013. No Expiration Date
- Machine Learning, Coursera, June 2013.
Publications and Patents Summary
Primary author of 27 papers in major peer-reviewed physics journals (such as Physics Review Letters, Physics Review D, etc.) and 15 papers published in proceedings of international conferences.
More than 50 talks are given at international conferences, workshops, and seminars.

Additional Experience

Training / Seminars
Teaching
- Object-oriented data analysis in high energy physics
- Computing in high energy physics
- Quantum eld theory

Fields of Expertise

Request Expert

Dev Tool:

Request: expert/machine-learning-applied-statistics-particle-physics-python-c
Matched Rewrite Rule: expert/([^/]+)(?:/([0-9]+))?/?$
Matched Rewrite Query: experts=machine-learning-applied-statistics-particle-physics-python-c&page=
Loaded Template: single-experts.php