TRIGGER ▸ portfolio.opened

Open to automation & integration work

Yurii Shchurko

Automation Engineer · n8n Specialist

I design and ship smart automations and AI ecosystems for business. Check out the architecture of my n8n workflows below.

See the workflows ↓ LinkedIn ↗ GitHub ↗

SET ▸ about

I turn manual, repetitive work into reliable automated workflows.

I design and ship smart automations and AI ecosystems for business. Check out the architecture of my n8n workflows below.

Remote · UTC+2

CODE ▸ stack

AUTOMATION
n8n Webhooks ETL Pipelines Human-in-the-Loop
AI
OpenAI Anthropic Pinecone pgvector LangChain MCP RAG
BACKEND
Node.js Express.js PostgreSQL Supabase REST
INFRA
Docker Cloudflare Git Linux

EXECUTE ▸ workflows

AI Personal OS — Telegram/Notion [Main Hub]

ProblemManaging personal productivity, habit tracking, and social interactions requires constant context-switching and manual data entry across multiple disjointed apps.
SolutionA Telegram-native AI operating system for personal productivity — a stateful multi-agent hub that routes voice and text messages to specialized sub-agents for task capture, day planning, social CRM, biometrics tracking, and semantic search. Built on a hub-and-spoke architecture with persistent Postgres memory and Supabase state management, each agent runs its own conversational lifecycle and commits structured data to Notion, Google Calendar, and Pinecone.
ResultUnified all life-management workflows into a single conversational interface, completely eliminating manual data entry.
n8nMulti-agent SystemRAGNL-to-JSON extractionWhisperPineconeHuman-in-the-LoopETLStateful sessionsPostgreSQLNotion APITelegram APIGoogle Calendar APIOpenAI API

API Gateway & State Manager

ProblemA multi-agent Telegram bot needs to gracefully handle context switching, user states, and multimodal inputs without mixing up conversational memories.
SolutionSyncs Telegram button interactions with Supabase DB to manage user states.
ResultEnsured zero context bleeding between sub-agents and provided a seamless, crash-free conversational routing experience.
n8nRoutingTelegram APIOpenAI APISupabasePostgreSQL

AI Personal CRM Agent (Vector-Powered)

ProblemKeeping track of networking history is tedious. Standard text matching fails to handle nicknames, typos, or new contacts efficiently in Notion databases.
SolutionA specialized agent that manages social interaction logs. It utilizes Pinecone semantic search to accurately resolve contact identities, dynamically fetches/generates Notion tags, and employs strict output parsers (Guardrails) to validate JSON payloads before database mutations.
ResultAutomated CRM logging with 100% data compliance, instantly resolving ambiguous contact names via vector entity linking.
n8nPineconeNotion APITelegram APIOpenAI APIPostgreSQLRAG

Context-Aware Biometrics Tracker

ProblemTracking daily health, habits, and productivity requires filling out complex forms manually, leading to friction and inconsistent data logging.
SolutionAn automated ETL pipeline that conducts dynamic, time-aware interviews. It implements a resilient "Two-Stage Commit" pattern: first backing up the raw narrative to Notion, then using an LLM to extract normalized properties (JSON) to update specific database fields.
ResultConverted unstructured daily reflections into structured biometric telemetry with zero data loss.
n8nSafe Data MutationTwo-Stage CommitNotion APITelegram APIOpenAI APIPostgreSQLETL

Autonomous Scheduling Agent

ProblemLLMs are notoriously unreliable at executing calendar updates because they fail to check for existing schedule conflicts or handle updates deterministically.
SolutionThis execution-focused agent translates high-level textual daily plans into precise CRUD operations (Create, Read, Update, Delete) within Google Calendar.
ResultEliminated double-booking and manual calendar management through a highly deterministic execution pipeline.
n8nOpenAILangChainAI AgentGoogle Calendar APINotion APIPostgreSQLJS

Notion → pgvector sync

ProblemEnterprise or personal knowledge bases are inaccessible to AI agents without a reliable, automated ingestion pipeline that keeps vector embeddings up to date.
SolutionAutomated n8n pipeline that syncs a Notion productivity database to a PostgreSQL vector store using OpenAI embeddings. Classifies items by type, resolves cross-entity relationships, and upserts idempotently — enabling semantic search over personal knowledge without manual intervention.
ResultProvided the foundational vector infrastructure, ensuring all conversational agents operate on the most recent, synchronized data.
n8nREST / HTTPVector embeddingspgvectorRAG infrastructureNotion APIPostgreSQLOpenAIJS

AGGREGATE ▸ stats

40+
workflows shipped
60+
integrations wired
120h
saved / month
1+
years automating

RESPOND ▸ contact

Let's automate something.

If you've got a repetitive process that should run itself, send it over. Most conversations start with a quick description of the manual steps.

Email me ↗ LinkedIn ↗ GitHub ↗ Download CV ↓