AI Architect / Team Lead

Software development company specializing in GovTech solutions and enterprise systems.

I designed and implemented a distributed, scalable web system for aggregating data into a centralized Data Lake in an air-gapped perimeter. The solution used Java 17, Spring Framework, Spring Cloud Data Flow, MongoDB, PostgreSQL, Elasticsearch, MinIO, Kafka, Redis, Angular, Gradle, Docker, Terraform, Kubernetes, ELK, GitLab CI/CD, and ArgoCD.

Architected an enterprise-grade Agentic AI orchestrating platform using Java, Spring AI, LangGraph, LLM, RAG, MinIO and Qdrant enabling multi-agent reasoning workflows for domain-specific tasks. Defined architectural standards for agent design, context management, and integration patterns across the organization.

Designed AI Voice Assistant with an automated OCR pipeline capable of processing millions of scanned documents. Defined the end-to-end architecture spanning speech recognition (Faster Whisper, ZipFormer), speech synthesis (Espeech, Silero), document understanding and classification (Qwen3-VL, PaddleOCR), LLM-based reasoning, and RAG retrieval using Java, Spring AI, Kafka, Temporal.io, MongoDB, Qdrant, SeaweedFS, Python, PyTorch, React, Label Studio, Terraform, Kubernetes, GitLab, ArgoCD.

Architected and developed a RAG platform for enterprise virtual AI assistants combining different types of RAG retrieval with a reproducible experiments, grounding checks, and an air-gapped deployment pipeline using Python 3.12, FastAPI, LightRAG, MSFT GraphRAG, Qdrant, OpenSearch, Neo4j, MongoDB, Redis, React, Docker.

I also implemented machine learning-based risk management systems, architected an enterprise Agentic AI platform, and designed an AI Voice Assistant with OCR and RAG capabilities. In addition, I authored the company's growth strategy and structured it using Enterprise Architecture principles aligned with TOGAF ADM.

Worked on a centralized e-Health system for the Republic of Belarus, with the customer being the Ministry of Healthcare.

I developed backend microservices focused on system security and compliance, and customized Keycloak with Java plugins to extend authentication functionality.

The stack included Java 11, Spring Framework, Keycloak, Kafka, ELK, FHIR R4, Maven, Git, and Docker.

Implemented various projects for customers from the USA, Europe, and Russia.

I worked on online presence, logistics, mobile banking, smart contracts analytics, computer vision, cloud framework, and IT spend management solutions. Across these projects I contributed to architecture, backend development, client-side modules, performance improvements, security, and data processing.

Technologies included Java, Spring Framework, AWS, MySQL, RabbitMQ, Angular, Keycloak, Gradle, Jenkins, Cassandra, Apache Spark, C++, Qt, and Apache ActiveMQ Artemis.

Designed and developed the client-side of an Electronic Document Management System using GWT for a Russian client.

I also developed a Spring-based backend for a Content Management System for a Norwegian client, adding virus scanning, format conversion, full-text search with Apache Lucene, and Dropbox integration.

Developed a web portal for testing and certification of personnel in the construction industry using Java.

I also created an easy-to-use UI/form component creator that improved interface design efficiency and reduced development time.

Master's degree in Computer Mathematics and System Analysis.