+ years in Software development with + years in Data science & Machine Learning, building data-driven solutions in various scenarios in science & industry. Highly proficient in the full cycle of development AI/ML-powered software. Armed with vast Data Science & Machine Learning, Leadership and Technical experience.
Member of AI Engineering team at Behavox - Industry leader of behavioral insights for Finance compliance solutions. Current role implies enabling Behavox platform and products with compliance/finance/regulations LLM models. Designing and implementing complete lifecycle for LLM models, including data science, data engineering, infrastructure and production developer roles.
Manager of ML Infrastructure team at Behavox - Industry leader of behavioral insights for Finance compliance solutions. Leading the team of 3+ engineers who delivered the new generation Behavox ML engine. Delivered architecture, design & implementation of the core microservices and tools to enable ML engineers, Data Scientists, Analysts with distributed, scalable and easy-to-use engine/tools to run ML applications on premises and clouds.
Senior Python Developer at Behavox - Industry leader of behavioral insights for Finance compliance solutions. Building new generation ML engine from scratch. Delivering design, architecture & implementation of the core microservice code to enable ML engineers, Data Scientists, Analysts with distributed, scalable and easy-to-use framework for running ML applications on premises and clouds.
Ideological inspirer, Team Lead, Chief Software Developer for DATASKAI - framework for rapid development of supervised ML solutions. Directing the team of 10+ people with various roles (python/java backend developers, frontend developers, UI/UX designer, tester, technical writer). Leading Technical/Architecture/ML in the side projects with industrial customers involving AI, ML in IoT Center of CDISE, Skoltech (done 4 projects involving planning, task formulation, planning, solving, delivering results).
As the team lead delivered with team two big PHM projects for biggest Russian airline companies. Software prototypes using ML techniques predict the Aircraft on the ground (AOG), faults and parts failure/repair events ahead of time in 2-7 days with 0.90-0.95 precision and 0.5-0.8 recall. Contributed a lot in planning, EDA, task formulation and formalization, prototyping, design and architecture, production coding, ML solution lifecycle, deployment, documentation and support
Built evaluation process for DSP and DMP models, done various dashboards and visualisations for continuous metrics review. Improved DSP/DMP models up to 15% in ROC AUC and logloss metrics. Reduced time of DMP assessment process from days to hours through developing various Python based tools.
Automated virtual environments setup on x86_64, powerPC architectures with AIX and Linux. Reduced time of virtual evironment setup from 2 days to 2-3 hours.
Delivered software for RedHat 6 based OS. Done a lot of tinkering with HA Clustering using RedHat. Built HA clusters with multi-platform nodes (x86_64, powerPC, SystemZ), master-slave DBMS replication using PostgreSQL, DB2, Oracle.
Participated in delivering of document management system for the municipal level of goverment.
First things first, on a day with nice weather you'll probably find me cycling. My target since august for this year (2023) is to do 300km per week until winter will stop me. Here is my profile on strava
Apart from the job I enjoy learning the neverending story of Artifical Intelligence. I like to read/research on books about AI, listen to podcasts, think about the future of AI, watch science fiction movies, particularly when some AI involved.
Examples of my contributions: lectures on AI, lecture on using singularity + snakemake for HPC computations, volunteering in the mlbinf project as Data Scientist and Infrastructure Engineer, building python libraries for cartesian genetics (docs) and genetic optimisation, giving a lecture on vowpal wabbit back in 2018
I admire TEX and use it for my documentation pages as well as slides (presentation example in tex).
Here are some examples of papers available on ResearchGate: