Independent AI engineer · Based in Germany
I build AI agents, knowledge search and workflow automation, connecting them to your databases, CRM and internal APIs. I can take a project from design to deployment or work alongside your engineering team.
Built an analytics platform that loads CRM data into a PostgreSQL warehouse. Users request metrics, reports and forecasts through a Telegram bot in natural language. Calculations rely on approved business metrics, and data queries run through a controlled execution layer.
Python · Airflow · PostgreSQL · MCP · Telegram Bot API
Led a pilot for support-ticket processing and response drafting. Designed an MCP integration layer for accessing Power BI, Amazon Redshift and Microsoft SQL Server, with Kafka-based event handling. Evaluated response quality and tool calls.
LangChain · MCP · Kafka · Power BI · Redshift
Developed workflows for extracting candidate information, matching CVs against vacancies, preparing shortlists and updating CRM records. Recruiters review external communication and consequential changes before they are sent.
Python · Qwen · n8n · CRM API
Contributed to a RAG system for internal DWH documentation. The solution indexes source materials and helps analysts locate specifications and data mappings with links to the underlying documents.
Python · Confluence · RAG · PostgreSQL
Designed integrations between internal systems, documenting REST API contracts, data structures and validation rules. Supported the engineering team through implementation, integration testing and issue resolution.
REST API · OpenAPI · SQL · PostgreSQL · Data Quality
Worked on moving reporting from Tableau and Power BI to Apache Superset. Analyzed data sources and metric definitions, developed PostgreSQL queries and checked figures against the original reporting systems.
PostgreSQL · SQL · Apache Superset · Tableau · Power BI
Proposed solution
Developed a concept for monitoring public procurement opportunities in urban electric transport. The proposed bot would track new tenders, review technical specifications and identify contracts aligned with the institute’s capabilities. It includes explainable suitability scores, a weekly Telegram digest and urgent deadline alerts. Staff remain responsible for the final bid decision.
Proposed stack · Python · PostgreSQL · aiogram · PyMuPDF · LLM API
Concept, not a deployed service. Access to the public procurement data still needs to be verified.
I build agents for handling requests and repetitive tasks. For example, an agent can find relevant information, draft a reply and send it to an employee for review.
Search across internal documents, Confluence and files. Responses link to source material and respect access permissions.
Custom MCP servers and Python/FastAPI backends that connect AI applications to CRM, ERP, databases and internal services.
Automate data extraction from PDFs and other documents. For example, compare supplier invoices with purchase orders and flag discrepancies for review.
We review your current process, available data and constraints. I identify what AI should and should not do.
I deliver an initial working version, test it against real workflows and review results with your team.
We add access controls, monitoring and documentation appropriate to the project, then deploy or hand over the solution.
My AI engineering work builds on a background in enterprise data systems. I have worked in systems analysis, data warehouse design, ETL pipelines and business reporting. This experience makes data sources, permissions and verifiable results central to how I design AI applications.
I graduated from the Faculty of Physics at Lomonosov Moscow State University and am now pursuing a master’s degree at Karlsruhe Institute of Technology (KIT).
Working stack · Python / FastAPI / PostgreSQL / Airflow / LangChain / MCP / Docker
A short description of the task and the systems involved is enough to begin. I’ll reply with a suggested next step.