About me
I turn data into autonomous AI solutions
A seasoned data expert with 15+ years of hands-on experience designing, building, and optimizing enterprise-scale analytical systems. I have command of the entire analytics stack — from data warehouses and ETL to databases and visualization — and lead full-cycle DWH projects from architecture to performance tuning, increasingly augmenting these systems with artificial intelligence in recent years. I don't just design systems on paper — I build them too, combining deep technical skill with architectural vision to turn complex data into clear business value.
In recent years, I have focused on integrating Artificial Intelligence into analytical and business systems. I design and deploy LLM-powered architectures with multi-agent orchestration that autonomously answer managers' questions, replace routine analyst tasks, monitor KPIs, and signal important deviations in real time.
I take a holistic approach to building analytical systems: I deliver complete solutions, not just proofs of concept. Every project is unique — designed from the ground up with respect to its business goals, budget, and infrastructure. I’m involved in each step — from code and infrastructure to optimization and visualization. This website itself runs on the data center I built, with all its code, deployment, and automation developed by me. I enjoy performance tuning, elegant data models, and creating systems that stay fast, stable, and transparent even under heavy load.
What I Do?
AI Agents & Orchestration
I design multi-agent systems: roles and permissions, task flows with branching and loops, multi-model routing, RAG, and guardrails. I build autonomous pipelines where agents drive research, development, testing, and deployment.
DevOps / MLOps
I build and maintain automated infrastructure that keeps data systems running reliably. From containerization and CI/CD to monitoring and security.
Data Warehousing
I document both your current system and the optimal architecture for your needs, including a detailed migration plan. Designs are confirmed through diagrams before any implementation begins.
Data Engineering
I create data warehouses optimized for your requirements. This could be a lightweight database on a single VM or a large-scale, distributed cluster architecture spanning multiple locations.
Data Visualization
I turn data into clear analytical reports from any source — from Excel and databases to APIs and real-time streams. I build interactive, user-friendly dashboards, accessible on mobile too.
Databases
I design, optimize, and maintain databases for your workload. From relational to NoSQL and columnar systems, I ensure reliability, scalability, and high performance with advanced query optimization.
Project Management
I create detailed project plans, find and interview top developers, lead the team, monitor progress, and make adjustments throughout development to ensure successful delivery.
Other
Broad technical expertise and a strong drive to learn new technologies enable me to quickly adapt to any challenge. Constantly exploring new approaches and tools to deliver effective solutions.
0 +
Enterprise DWH built
0 TB+
Volume of data processed
0 min
Analytical data delay
0 +
AI agents orchestrated
AI Agents & Orchestration
Orchestration engines: custom engine, CrewAI, AutoGen, LangGraph
85%
LLM & routing: OpenAI, Anthropic, Google, Mistral, Llama, DeepSeek; OpenRouter, LiteLLM
80%
RAG & memory: Qdrant, Weaviate, FAISS
74%
Guardrails & monitoring: access levels, evals, tokens, latency
75%
DevOps / MLOps
Containers: Kubernetes, Docker
85%
CI/CD: Git, Deploy, Auto-tests
90%
Networks: VPN, Subnets, Firewall
75%
Hardware: CPU, RAM, Storage
70%
MLOps: model deployment, monitoring, retraining, guardrails
70%
Data Warehousing
Standards: Data Vault, Anchor modeling, EAV, Star, Snowflake, 1/2/3NF, Unstructured data
95%
Approaches: Bill Inmon, Ralph Kimball, Hybrid approach
90%
Scheduling: Argo, Prefect, Airflow, Jenkins, Rundeck
85%
Modeling: SAP PowerDesigner, MS SSMS, MySQL Workbench (E/R, IDEF1X, Barker)
80%
Data Engineering
Querying: SQL, PL/SQL, T-SQL
95%
Programming: Python, Java, Scala, R; Frameworks: Spark, PySpark, SparkR, Beam, Storm, Flink
85%
ETL Tools: Informatica, Pentaho, SSIS, AWS Glue, Alooma, Stitch, Zapier
90%
ELT and data loaders: DBT, Trafodion ODB
80%
Streaming: NiFi, Debezium, Kafka connectors
75%
Data Visualization
Tools: Superset, Tableau, Power BI, SSRS, QlikView, Metabase
90%
Cloud-based: Looker, Klipfolio
80%
Monitoring: Grafana, Zabbix
85%
Programming: D3.js, Chart.js, Plot
75%
Databases
Columnar: Vertica, Snowflake, Greenplum, Clickhouse, Exadata, Redshift, BigQuery
95%
Row-oriented: MySQL, PostgreSQL, MS SQL Server, Oracle
90%
Multidimensional: SSAS, Mondrian, Oracle BI
85%
NoSQL: MongoDB, Cassandra
80%
Project Management
PM Tools: JIRA, Redmine, Trello, Asana, MS Project
85%
Task Management: Structuring and assigning tasks to the team
90%
Hiring: Interviewing and hiring strong specialists
85%
Other
Front-end: NextJS, HTML, CSS
60%
Back-end: Python, NestJS
75%
Founder, Agentic Systems Architect
2025-Present (Orchestra (AI Organization))
My personal achievements:
• Built a custom-made AI organization where teams of agents and humans autonomously handle research, development, testing, and deployment. • Implemented a declarative flow-as-code description with online monitoring and agent permission control. • Implemented a loop that auto-selects the best-fit model for each role.
My responsibilities:
• Design and build the "artificial office" engine: AI agents are full-fledged employees on par with humans, with the same authority and responsibilities, and they communicate with each other. • Own orchestration as code: a declarative configuration defines roles, who communicates with whom, and the task flow — the engine enforces this and generates the agent communication logic itself. • Lead complex flow with branching and loops. For example: a task goes in parallel to development, marketing (newsletter), and accounting; the team lead decomposes it among developers, who communicate and hand parts off to the architect for review; technical writers document it, DevOps publishes and triggers the newsletter; an analyst then accepts the work based on analytical-reporting data (e.g., link clicks). • Own online monitoring and statistics: which agents and employees are involved, how many tasks each completed, how much (money and time) each task and each employee cost, plus system load. • Ensure the access-control system: each agent has its own permissions, and a separate control layer enforces that the agent stays within its bounds. • Lead the loop that tests and selects the best-fit model for each role. • Own autonomous execution of tasks: development, marketing, reporting, accounting, and others.
Project technologies:
• Orchestration: custom-built engine, flow as code (DSL); evaluated CrewAI, AutoGen, LangGraph before building the custom engine • Models: LLM selection and routing (OpenAI, Anthropic, Mistral, Llama), RAG (Qdrant, Weaviate) • Control: agent access levels, online monitoring • Infrastructure: Python, Kubernetes, Argo Workflows
Founder & Tech Lead
2023-Present (Strive (Data analyzing))
My personal achievements:
• Launched a fully operational automated trading system that generates real revenue. • Made the system exchange-agnostic — a single pipeline for the order book and trades of any venue. • Built a continuous retraining loop with a deployment gate: a model doesn't ship to production until it passes validation. • Integrated AI modules for analyzing unstructured financial data and news via RAG. • Launched autonomous AI pipelines combining trading logic and generative insights.
My responsibilities:
• Own the architecture of the trading-analytics system and the trading algorithms — designed from scratch. • Ensure exchange-agnosticism: connect to any venue (only the order book and trade stream are needed), normalize formats across exchanges. • Lead the execution layer: order routing, slippage and venue-fee accounting, profit estimation net of costs — on live trading. • Own the pipeline for continuous background model retraining on a cluster; a model only ships to production through a deployment gate (validation/falsification). • Develop ML models (PyTorch/scikit-learn) for pattern and anomaly recognition on non-stationary time series, and LLMs with RAG for sentiment evaluation and news analysis. • Lead multi-agent orchestration of research, development, testing, and deployment.
Project technologies:
• Data Sources: exchange APIs (order book and trades), external APIs • ETL & ML: Python, Argo Workflows, Kafka; PyTorch, scikit-learn (retraining) • DWH: HPE Vertica, Kimball model • Visualization: Superset, Chart.js
Head of Data
2020-2024 (Sravni.ru (Web))
My personal achievements:
• Built the entire system from scratch, ensuring stable operation. • Designed end-to-end architecture from products to analytics users. • Negotiated optimal pricing for all solution components. • Led migration from legacy slow DWH to new high-performance system. • Developed and enforced company-wide data workflow policies. • Deployed an internal AI assistant that automated analytical requests using RAG and prompt-engineering. • Reduced report preparation time by over 50% and accelerated decision-making company-wide.
My responsibilities:
• Selected and deployed servers and software solutions across the company. • Owned company-wide data flow architecture, with a focus on DWH. • Defined and enforced data modeling standards and rules. • Monitored and prevented personal data leaks, ensuring data security compliance. • Researched and implemented LLM-based assistants for the analytics team: automating data research, insight generation, and documentation. • Deployed ML anomaly-detection modules and LLM-based SQL suggestions, boosting the speed and accuracy of analytics.
Project technologies:
• Data Sources: MSSQL, PostgreSQL, MySQL, MongoDB, External APIs, RabbitMQ, Kafka, CSV, JSON • ETL: Python, TeamCity, Prefect, Kafka, Debezium • DWH: Snowflake, Kimball model • Visualization: PowerBI, Snowsight, Sigma, Grafana
DWH Architect
2016-2020 (SuperJob (Web))
My personal achievements:
• Developed a stable analytical system from scratch. • Optimized budget and resources by sizing licenses, disk space, and designing a container-based test environment on 2 hosts. • Developed ETL pipelines using Pentaho and Python. • Built an autonomous auto-testing system requiring no manual intervention. • Delivered a Big Data solution covering the full range of analytical tasks. • Designed a Data Vault model enabling fast, high-quality business data, including ML scoring, ranking, classification, and MDM. • Collaborated with analysts to release real-time dashboards on centralized storage. • Implemented a 3NF metadata model for centralized data flow coordination and quality monitoring.
My responsibilities:
• Owned the fully functional analytical system: built it from the ground up and maintained it. • Selected and implemented servers and software solutions to optimize performance and cost. • Led the Data Vault model and ETL metadata in PowerDesigner. • Administered an HP Vertica cluster on *nix servers, expanded the cluster, and balanced loads to maintain minimal data latency, ensuring no user requests queued even at peak delays of 40–50 minutes. • Owned analytical ML procedures: on-the-fly user-type classification, scoring/ranking models.
Project technologies:
• Data Sources: MariaDB, MongoDB, CSV, JSON, External APIs • ETL: Pentaho DI, Python, Jenkins, Airflow • DWH: Vertica+Clickhouse, Data Vault model • Visualization: Tableau, Grafana
Lead ETL Developer
2014-2016 (Svyaznoy (Retail))
My personal achievements:
• Created a unique adaptive data model for a Big Data warehouse using a cluster-based approach. • Developed Python-based automated tests covering all ETL processes. • Designed and implemented a Data Quality (DQ) system that caught most bugs before deployment, ensuring high reliability of ETL processes.
My responsibilities:
• Led the end-to-end development of a large-scale data warehouse project. • Communicated directly with customers to gather requirements and feedback. • Built and implemented data models. • Planned and coordinated team activities, ensuring efficient workflow. • Refactored and validated the majority of ETL processes.
Project technologies:
• Data Sources: Hadoop, Oracle, MSSQL, PostgreSQL, MySQL • ETL: Pentaho DI, Python, Sqoop, Kafka, Oozie • DWH: Hive, Vertica, EAV
Senior DWH Developer
2012-2014 (Asseco group, R-Style SoftLab (IT consulting))
My personal achievements:
• Integrated and completed the Data Warehouse (DWH) system in several top-20 Russian banks. • Developed management reporting and IFRS-compliant reports. • Delivered the final data visualization for end users.
My responsibilities:
• Developed large-scale DWH projects from scratch for major Russian banks. • Delivered end-to-end solutions: sales support, analytics, ETL, DWH, reporting, and client acceptance. • Implemented Oracle DWH (RSDH) and integrated it with bank systems. • Built ETL pipelines, OLAP reporting, and data validation processes. • Conducted operational training and certification programs for new team members. • Supported analysts with Oracle and RSDH, ensuring reliable reporting for banking operations.
Project technologies:
• Data Sources: Oracle, MSSQL, PostgreSQL, MySQL • ETL: Informatica powercenter • DWH: Oracle Exadata, 3NF Model
Data Quality Expert
2011-2012 (TNS Global (Research))
My personal achievements:
• Designed real-time data quality reports and alerts, highlighting key system issues instantly. • Created simulation schedules for a web-monitoring robot, modeling user behavior by age, location, and gender. • Developed and launched a custom Reporting Service in-house, enabling accurate and timely analytics.
My responsibilities:
• Owned the OLAP warehouse: built it from scratch and maintained it, ensuring high-speed updates of large cubes. • Led automation of data aggregation and reporting — accuracy and timeliness. • Optimized SQL procedures, jobs, and triggers for maximum performance. • Implemented SSAS and SSRS solutions, streamlining statistical reporting and analytics.
Project technologies:
• MSSQL, SSMS, SSIS, SSAS
Database Migration Specialist
2009-2010 (Allianz, Rosno (Insurance))
My personal achievements:
• Successfully migrated complex datasets from multiple systems into SAP with full data integrity. • Optimized SAP module performance, accelerating operational reporting. • Managed the entire migration workflow from extraction to validation.
My responsibilities:
• Created technical specifications and calculation algorithms for SAP modules. • Tested and optimized SAP module performance. • Exported and corrected data from INFIN, CIS, and other corporate systems to ensure smooth migration into SAP. • Prepared and validated data for migration, ensuring both accuracy and completeness.
Project technologies:
• SAP, Oracle, Microsoft Office, SQL, VBA
Actual
Autonomous AI Organization
Orchestra (personal project)
Orchestra is an artificial organization integrated into real business processes: employees — AI agents and humans — work side by side, with equal authority and responsibilities. It describes roles, authority and responsibilities, the tasks of different teams, and the processes tasks follow through the flow. A personal project: the orchestration engine and all its logic were written by me.
Project technologies:
• Orchestration: custom-built engine, flow as code (DSL) • Models: LLM selection and routing (OpenAI, Anthropic, Mistral, Llama), RAG (Qdrant, Weaviate) • Control: agent access levels, online monitoring • Infrastructure: Python, Kubernetes, Argo Workflows • Hosting: My own data center
Trading automation system
Strive (Seoul, South Korea)
Strive is an automated trading system (bot) that works with any exchange order book. It analyzes the order book and trade flow, and forecasts future quotes based on machine learning and news sentiment analysis. Founder and sole engineer: designed and built the system end to end.
Project technologies:
• Data Sources: exchange APIs (order book and trades), external APIs • ETL & ML: Python, Argo Workflows, Kafka; PyTorch, scikit-learn (retraining) • DWH: HPE Vertica, Kimball model • Visualization: Superset, Chart.js • Hosting: My own data center
Data analyzing API layer
Totle (Tel Aviv, Israel)
Totle is a high-performance real-time Ethereum data analytics layer and decentralized exchange (DEX) aggregator: routing orders across liquidity pools for the best execution price, accounting for slippage. The analytics layer — the DWH model and the order-routing API with execution estimates — was built by me.
Project technologies:
• Data Sources: Blockchain • ETL: Node.js • DWH: Snowflake, 3NF model • Visualization: d3.js, Node.js API • Hosting: AWS
Distributed data center
OpenHub (Davao, Philippines)
OpenHub is a distributed data center with sites across several countries. It combines GPU farms and high-performance analytical servers into a single network, providing compute infrastructure for analytical and AI workloads. The entire data center network and infrastructure were designed and assembled by me.
Project technologies:
• Hardware: Supermicro, Intel, RAID, Cisco, Mikrotik • Software: Windows Hyper-V, Ubuntu • Visualization: Grafana, UptimeRobot • Hosting: My own data center
Archived
Data streaming system
Sravni.ru (Moscow, RF)
Sravni.ru is a company-wide real-time data streaming system into the DWH (CDC), built on a Kubernetes architecture with a data mesh approach for microservices. The company's entire analytical system and all its internal data flows were designed, documented, and developed by me.
Project technologies:
• Data Sources: MSSQL, PostgreSQL, MySQL, MongoDB, External APIs, RabbitMQ, Kafka, CSV, JSON • ETL: Python, TeamCity, Prefect-Kubernetes, Kafka, Debezium • DWH: Snowflake, Kimball model • Visualization: PowerBI, Snowsight, Superset, Grafana • Hosting: 100% cloud managed
Customer analytics
PharmaKey (Moscow, RF)
PharmaKey is a pharmacy analytics system for 10+ clients, each processing over a billion transactions, with nationwide production and sales insights. The whole system and the migration to Vertica were delivered by me — ETL, warehouse, and visualization.
Project technologies:
• Portal backend: PHP, JavaScript, MySql Data Sources: MySQL, CSV, External APIs ETL: Migrated from Rundeck and Bash to Pentaho DI, Python, Airflow • DWH: Migrated from MySql and Redshift to Vertica, Snowflake model • Visualization: Tableau, Metabase • Hosting: Migrated from AWS and Hetzner to Selectel dedicated+vmWare
Enterprise analytics system
SuperJob (Moscow, RF)
SuperJob is an enterprise analytics system: a DWH model and analytics layered on more than 20 data sources, handling on the order of a billion events per week. The entire analytical system was developed by me, from scratch — ETL, warehouse, and visualization.
Project technologies:
• Data Sources: MariaDB, MongoDB, CSV, JSON, External APIs • ETL: Pentaho DI, Python, Jenkins, Airflow • DWH: Vertica+Clickhouse, real-time ods + Data Vault model + near real-time star • Visualization: Tableau, Grafana • Hosting: Enterprise dedicated
Customer analytics
PetroSoft (Pittsburgh, USA)
PetroSoft is a customer analytics system for a network of gas stations in the USA and Canada: real-time dashboards handling on the order of a billion transactions per week. The whole system was built by me — ETL, warehouse, and real-time dashboards.
Project technologies:
• Data Sources: MySQL, Enterprise Bus • ETL: Pentaho DI, Python, Jenkins • DWH: Vertica, Snowflake model • Visualization: Tableau, JavaScript • Hosting: Enterprise dedicated
E-commerce sellers analytics
Rafferi (San Francisco, USA)
Rafferi is an analytics system for e-commerce sellers: integrating data from Amazon, eBay, and Walmart into a single warehouse with dashboards and sales reporting. The entire analytics — from source integration and ETL to the warehouse and dashboards — was designed and developed by me.
Project technologies:
• Data Sources: AWS API, MWS API, eBay API • ETL: Pentaho DI, Java • DWH: Vertica, Star model • Visualization: d3.js, Node.js API • Hosting: My own data center
Education
Lomonosov’s Moscow State University
Mathematical methods in economics
2014
Russia's oldest and one of its leading universities, consistently featured in global rankings. Admitted to a highly competitive state-funded place.
Moscow State Technical University n.a. N.E.Bauman
Object-oriented programming
2009
One of Russia's leading technical universities, with a strong engineering tradition. Admitted without entrance exams after winning the Moscow Physics and Mathematics Olympiad.
MITx Massachusetts Institute of Technology
Introduction to Computer Science and Programming
2014
MITx online course (edX) in computer science and programming, completed with a final assessment.