Oleg H.

I build payment systems and the AI agents that run on top of them. I am a senior backend engineer with 9+ years in Java and Spring Boot, including 4+ years on the billing and payments platform behind a marketplace moving about $4B a year. PAYMENTS AND FINTECH I deliver provider integrations end to end: Adyen, Braintree, Payoneer, PayPal, Venmo, Wise, Wells Fargo, Bank of America and Forter, on a platform operating across 180+ countries. My work covers card processing, wallets, payouts, bank transfers, fraud screening and reconciliation over REST, SFTP and webhooks. I have also integrated Stripe, FastSpring and bank vPOS in client products. Representative work: - Migrated a major payout provider from a legacy protocol to REST with OAuth 2.0. The migration covered payouts, mass payouts, payee registration, card release and reconciliation. I shipped it to 100% behind feature flags with no customer-facing incidents. I built provider emulators to test it without vendor sandboxes. - Designed a card-processing service on a provider's GraphQL API covering client tokens, payment methods, network tokens, cryptograms, authorization, capture, refunds, cancellations and 3D Secure. - Built a forwarding layer that routes network tokens between two providers, enabling non-US card acceptance. I rolled it out in stages and negotiated timeout SLAs directly with the vendor. - Authored the design for splitting a payments monolith into nine provider-specific microservices. Domain coverage: 3DS2, network tokenization, cryptograms, authorization and capture, refunds, chargebacks, reconciliation through APIs, webhooks and SFTP files, idempotency and double-charge prevention, mass payouts, ACH returns, fraud screening, PCI-aware design. AI AGENTS AND LLM SYSTEMS I build AI that runs inside products with real users and revenue. I built an agentic system that investigates failed payment transactions on its own, reasoning across logs, dashboards, source code and incident tooling to cut manual triage. I built it with Python, LangChain and multi-agent orchestration. It runs in a high-stakes payments environment, where being wrong has a cost. I also built and run Memabooks, a profitable AI commerce product that turns a child's photo into a personalized illustrated storybook. Over 10,000 books created so far, printed on demand and shipped worldwide, with customers in eight countries. I developed a multi-stage, identity-preserving image generation pipeline covering face detection, background removal, personalized character generation, skin-tone matching and character consistency across a complete book. I work with diffusion models, fine-tuned LoRA adapters and serverless GPU inference, tuning for quality, latency and cost. I also built the Node/NestJS backend, the Angular front end, print-on-demand fulfilment, and the payment stack for a market with limited access to major global processors: bank vPOS, 3DS2, Apple Pay, idempotency and double-charge prevention, and abandoned-payment recovery. AI stack: LangChain, LangGraph, LlamaIndex, multi-agent systems, RAG and multimodal RAG, vector databases (FAISS, Chroma), Model Context Protocol, OpenAI, Gemini, Azure AI. BACKEND ENGINEERING AND RANGE My core stack is Java 17 and Spring Boot. I also have production experience with Node.js, NestJS, TypeScript, Python, PostgreSQL, MongoDB, AWS, Docker, Kubernetes, REST, GraphQL, microservices and event-driven architecture. I was the sole engineer responsible for Geoapify's product surface for 17 months, working directly with the product owner across 14 repositories. I built a geocoding SaaS end to end, covering file upload, worker-based job processing, object storage and card payments. I maintained their public TypeScript SDKs for MapLibre GL, Leaflet and GeoJSON, which see about 55K npm downloads a month, and owned their public API playgrounds and OpenAPI specs. Testing is part of my delivery process. On one project, I built 260 Playwright end-to-end tests from scratch in about six days. WHAT I CAN HELP WITH - Payment provider and fintech integrations - Java and Spring Boot backend development - Backend architecture and monolith-to-microservices migrations - AI agents, RAG and LLM integration - API design, reliability, observability and production troubleshooting CREDENTIALS - Oracle Certified Professional: Java SE 11 Developer - IBM RAG and Agentic AI Professional Certificate - Google Cloud Generative AI Leader - Microsoft Azure AI Fundamentals - BSc in Computer Science, American University of Armenia - Former Java lecturer, Russian-Armenian University My longest engagement here ran 19 months and 2,700+ hours. I am open to long-term, full-time and contract-to-hire engagements. Tell me what you need built or what is currently breaking. I will tell you honestly whether I am the right engineer for it.