I design reliable business workflows with n8n, AI/LLMs, APIs, CRM systems, and data tools.
My projects focus on the parts that make automation maintainable: data contracts, deterministic routing, controlled AI outputs, failure paths, operational visibility, testing, and handover-ready documentation.
Portfolio · Русское резюме · Telegram
- AI-assisted lead qualification and request triage with validated model outputs
- API and webhook workflows with authentication, idempotency, and explicit responses
- RAG systems with separate ingestion and query boundaries
- CRM, Telegram, Gmail, Google Sheets, Notion, and external REST API integrations
- Scheduled monitoring, data normalization, operator alerts, and reporting
- Documentation and test scenarios that make workflow behavior reviewable
| Case study | Business use case | Engineering signal | Stack |
|---|---|---|---|
| Barbershop Booking API | Validate, store, deduplicate, and notify on bookings | Modular workflows, practical idempotency, async notifications, centralized errors | n8n, Webhooks, Google Sheets, Telegram, Gmail |
| RAG Telegram FAQ Bot | Answer support questions from a controlled knowledge base | Separate ingestion/query paths and explicit retrieval limitations | n8n, Supabase, pgvector, Hugging Face, Groq, Telegram |
| AI Lead Processing Pipeline | Qualify and route inbound sales leads | Validation before AI, parsed model contract, deterministic routing | n8n, Groq, CRM API, Telegram, Gmail, Google Sheets |
| Competitor Price Monitoring | Detect changes on public product pages | Deterministic comparison, explicit extraction failures, AI-only enrichment | n8n, JavaScript, Groq, Google Sheets, Telegram |
The complete project index, exports, examples, screenshots, test scenarios, and limitations are in the n8n Automation & Integration Portfolio.
Business process
→ boundaries and data contracts
→ deterministic workflow + selective AI
→ validation and failure paths
→ observability and testing
→ handover-ready documentation
I use AI where semantic judgment adds value and keep validation, permissions, routing, and state transitions deterministic whenever possible.
I treat a successful demo as the beginning of review: retries, idempotency, partial failures, concurrency, personal-data handling, and deployment constraints still need explicit decisions.
Automation: n8n, webhooks, scheduled workflows, sub-workflows, Error Workflows
AI & data: LLM agents, structured outputs, RAG, embeddings, Supabase / pgvector, Google Sheets
Integrations: REST APIs, CRM systems, Telegram Bot API, Gmail, Google Drive, Notion
Code & tools: JavaScript, JSON, Git, GitHub, Docker, VS Code
Current focus: orchestration, state, memory, reliable tool use, and production deployment.
Open to freelance automation projects and AI automation / integration roles.
Working languages: Russian and English.
- Telegram: @Andyyy_Randyyy
- GitHub: Andy-randy
- Portfolio: n8n-portfolio
Я проектирую надёжные бизнес-процессы на базе n8n, AI/LLM, API, CRM и систем хранения данных. Основной фокус — не отдельный «бот», а весь рабочий контур: входные данные, границы компонентов, маршрутизация, ошибки, наблюдаемость и передача проекта.
В портфолио лучше начать с Barbershop Booking API, RAG Telegram FAQ Bot, AI Lead Processing Pipeline и Competitor Price Monitoring. В README проектов есть схемы, workflow exports, примеры контрактов, тестовые сценарии, известные ограничения и план production hardening.
Открыта к freelance-проектам по автоматизации и ролям в AI automation / integration. Связаться можно через Telegram.