Oleg B.

I build production B2B SaaS with AI at the core. One engineer who owns the full stack: frontend, backend, infra, and the AI itself. Selected production work: VendorFox (vendorfox.ai): B2B SaaS for Managed Service Providers, tracking hardware lifecycle, firmware, and security advisories across 48+ vendors. Built end-to-end by me. Next.js + TypeScript frontend, FastAPI backend on Cloud Run, Firestore multi-tenant, Firebase Auth, Stripe billing across four tiers. The AI layer is DAG-based LLM orchestration with structured outputs and an aggressive dedup and caching design that keeps API spend flat as data volume grows. Live, paying customers. DataAbyss (app.dataabyss.ai): earlier production work. OSINT intelligence platform used by government and national security professionals. Elasticsearch across billions of entity connections, RAG pipelines, LLM-powered chat interface. Gold Stevie Award for Achievement in Science or Technology, 2023 American Business Awards. What I build: Production LLM orchestration (structured outputs, retry logic, cost control that keeps API bills predictable), RAG pipelines with real retrieval tuning, vector search on pgvector and Pinecone, multi-tenant SaaS with Stripe billing and Firebase Auth, distributed agent workflows with concurrency control and fault tolerance. Stack: Next.js, TypeScript, FastAPI, Python, Node.js, PostgreSQL, Elasticsearch, Firebase, AWS, GCP, Cloud Run, Docker, LangChain, Claude, OpenAI, Stripe. Background: First-class Physics degree, Open University UK (4.0 US GPA equivalent)