AI Workflow Architect — I connect business tools to LLMs and make the whole stack run without babysitting.
Previously at CoinTracker, where I built production AI systems tying together Slack, Notion, GitHub, Linear, Guru, and Canny — triage bots, auto-routing pipelines, workflow automation that actually stayed running.
Now I build open-source templates and take on consulting work for teams that want real AI integrations, not demos.
All templates use pydantic-ai for structured LLM outputs and FastAPI for the webhook/API layer. Drop one into your project and have a working AI integration in under 10 minutes.
| # | Template | What It Does |
|---|---|---|
| 01 | Slack → Notion Automation | Classifies Slack messages, auto-creates Notion entries for tasks/decisions/blockers |
| 02 | GitHub Issue → Linear Triage | Reads new issues, assigns priority + story points, creates Linear tickets automatically |
| 03 | Multi-LLM Cost Optimizer | Routes prompts to the cheapest model that handles them — Claude, GPT-4o, Groq |
| 04 | Web Scraper + Semantic Search | Scrapes any URL, extracts structured content, keyword + AI-ranked search |
| 05 | Prompt Engineering Runbook | 5 agents: SWOT analyzer, social post generator, code reviewer, summarizer, decision framework |
| 06 | Email → Linear Auto-Triage | Classifies customer emails, sets priority, creates Linear issues automatically with Slack alerts for critical tickets |
All templates are available individually or as a bundle on Gumroad.
APIs I work with regularly: Anthropic Claude, OpenAI, Slack, Notion, GitHub, Linear, Guru, Canny
Infrastructure: FastAPI, uvicorn, httpx, Playwright, pydantic-ai, litellm
I take on a small number of consulting engagements for teams that want AI workflows that actually stick.
What I do:
- Connect your existing tools (Slack, Notion, GitHub, Linear, etc.) to LLM pipelines
- Build triage systems, auto-routing, and workflow automation
- Structure your prompts so outputs are always machine-readable (no regex parsing)
- Audit existing AI setups for reliability and cost
Typical engagement: 1-4 weeks, fixed scope, clear deliverables.
To start a conversation: cw.allen0083@gmail.com or book a discovery call (Calendly link — coming soon)
Building in public — all templates are open source.