I'm a web developer focused on building internal tools, platform integrations, and practical systems for operational teams.

My work often sits between product engineering and support operations: connecting tools like Zendesk, Slack, Jira, and AWS so teams can work with better context and fewer manual steps.

More recently, I've been exploring how language models and agent-based workflows can support incident response, product decisions, and other operational processes—particularly where AI can assist a human decision rather than replace it.

An incident-response prototype that brings observability and support signals into a coordinated workflow.

It explores:

• MCP-based tool integration
• Structured model outputs
• AI-assisted incident triage
• Customer-impact analysis
• Human review and approval
• Deterministic workflow coordination

A Slack-native workflow for evaluating product feedback and recording decisions.

It explores how specialized perspectives, structured recommendations, and human approval can work together inside an existing team workflow.

A customer-observability system that connects product telemetry with account and revenue context.

Sybil uses deterministic rules and statistical baselines—not an LLM—to identify unusual behavior and help prioritize affected accounts.

Languages and frameworks

TypeScript JavaScript React Next.js Node.js

Platforms and infrastructure

AWS PostgreSQL Terraform Vercel Firebase

Integrations and workflows

REST APIs Slack Zendesk Jira Model Context Protocol

I’m especially interested in systems that:

• turn fragmented signals into usable context;
• fit into the tools people already use;
• keep important decisions inspectable;
• separate AI responsibilities from deterministic application logic;
• give humans control over consequential actions.

You can see more of my work at kateives.com.