Oleh – Golang, Golang REST API, AI agent development

A legal tech SaaS platform for structured data extraction from M&A agreements, serving large law firms globally. Core capabilities include high-accuracy document parsing, semantic search over extracted data, customizable views with advanced filtering and charting, and an AI chat interface. The platform handles enterprise-scale document workflows with strict per-customer access control.

• Restructured storage-layer data models in a large-scale refactoring effort, enabling fine-grained enterprise access control across customer datasets and workflows
• Reduced tokenization failures by 90% through search vector optimization, improving platform reliability and minimizing downstream pipeline failures
• Accelerated search performance by 30x by optimizing vector retrieval and launching metadata-aware semantic search
• Built the Views (Portals) feature — a customer-facing data framework with configurable charting, advanced filtering, and role-based access controls
• Created Grafana dashboards for production observability across the GCP/Kubernetes infrastructure
• Built internal AI engineering tooling using Claude Code, including a fully agentic bug-resolution pipeline that reproduced issues locally, opened fix PRs, verified results, and produced code-review walkthroughs

A PCI-compliant fast checkout platform integrated directly into merchant e-commerce flows, handling payment credential collection, customer authentication against card networks (Visa, Discover), and end-to-end payment processing. Served merchants and their customers at high availability across five years of scaling and security-critical growth.

• Owned security architecture including encryption-at-rest, tokenization, and network isolation — contributing to zero security incidents over five years
• Designed and deployed HashiCorp Vault infrastructure for secret management and cryptographic operations across the platform
• Drove PCI DSS compliance across four consecutive successful audits; architected a secure card storage deployment using Terraform + Vault with a CircleCI-based infrastructure change pipeline for auditable, mission-critical updates
• Established the foundational Kubernetes cloud platform with distributed tracing, authentication, and integrity monitoring; provisioned all cluster infrastructure via Terraform
• Built the core Golang monolith spanning authentication, payments, and orders, scaling it into a multi-service payment platform
• Built API integrations with Visa and Discover; defined the card-tokenization blueprint for all future card-network integrations
• Architected ACH bank-payment integration including secure customer banking credential storage
• Introduced a shared internal Golang library of best practices and reusable service frameworks, adopted as the engineering standard across teams; recruited and mentored 12 engineers
• Set up CI/CD pipelines for 8 core microservices with Puppeteer/Headless Chrome automation scripts; reduced core service latency by 1.5–2x via load testing and optimization
• Maintained 99.95%+ SLA; migrated ~1M data records to PostgreSQL as part of a company merger integration

Independent quantitative trading research project — building a backtesting platform and running algorithmic strategy research on high-frequency financial data using Python and ML tooling.

• Re-engineered a backtesting platform, achieving 12x throughput improvement on quantitative strategy evaluation
• Conducted quantitative trading research using Python, Jupyter, and TensorFlow/Keras
• Built ETL pipelines processing high-frequency financial data into BigQuery

Event-driven data pipeline services for IoT device data, focused on accuracy, throughput, and quantitative analysis infrastructure.

• Built event-driven Golang and Kafka services that raised device-data accuracy by 45%.

A workforce-management SaaS platform, with a focus on service reliability and automated testing discipline.

• Built REST and JSON-RPC Golang microservices for a workforce-management platform serving enterprise clients.
• Raised automated test coverage to 80%, cutting production incidents by 3x.