I’m Sergey Yashchuk, a software engineer and AI enthusiast. I build tools that connect AI clients, structured data and automation workflows.
My focus is MCP servers, automation pipelines with n8n, and developer tools. I’m also exploring reusable agent skills: clear instructions, tool use and checks that make a workflow easier to repeat.
Explore selected work · Discuss a project
A lead pipeline controlled through an MCP client.
A local TypeScript server with MCP tools, resources and a prompt for managing lead records, follow-ups and CSV files. An optional webhook connects it to n8n.
Example: add a lead → record a qualification score → schedule a follow-up → preview an n8n event. Walk through the scenario.
- Try it locally: fictional seed data; the local server needs no API keys.
- Inspect the workflow: stdio MCP calls and HTTP webhook requests are covered by tests.
- See the limits: the recorded webhook call is a dry run; the example email node is disabled.
Source & setup · Watch the demo · Tests
n8n workflow example — filters followup_email events and maps the event, lead ID and payload. It is inactive, with a disabled email placeholder; this starter does not include a full validation/review path.
- Android Force 120Hz — a Kotlin accessibility utility for LTPO displays, with a demo and published APKs. Device compatibility and battery use remain practical trade-offs.
- Simple Outbound Webhooks — an event-to-HTTPS integration for WordPress, with optional HMAC signatures, a delivery log and a Docker test harness.
I’m interested in collaborating on MCP integrations, AI-assisted workflows and n8n pipelines. I’m also exploring reusable agent skills. A good starting point is one concrete task, a non-sensitive sample input and a result we can check.
For a small agency assignment, I offer one scoped n8n validation / mapping component, with sample checks and a documented handover.