# Franklin Christuraj

> Building agents that survive contact with a real business. The production reality layer: agentic governance, shared state, and hard-coded human checkpoints.

> Value Engineer at Make (Celonis). Enterprise AI architecture for multi-agent systems that survive production, not just the demo.

Franklin Christuraj (Frank About AI) publishes work, offers (Embedded Build Sprints and systems audits), a build log of engineering decisions, and a way to start a conversation. The site is fully static and public.

Full multi-agent orchestration frameworks, project corpus, and build log: https://frankaboutai.com/llms-full.txt

## What this site is about

- Agentic architecture: multi-agent systems designed around a single source of state, with human-in-the-loop checkpoints where the cost of being wrong is real.
- Embedded Build Sprints: hands-on enablement where client teams ship working automations inside their own governance. 600+ professionals trained.
- Multi-agent orchestration: production patterns for coordinating several agents (shared state, typed edges, failure-mode analysis).
- Agentic governance and the production reality layer: acceptable-use policy, human review steps, audit trails. Making agentic systems legible to an engineer and a CFO in the same frame.
- Build log: dated decision entries (constraint, options, chose) plus unresolved open problems.

## Who it is for

Engineering leaders, innovation teams, and operators who need agents that are reliable in production, not just in a demo deck.

## Pages

- [Home / overview](https://frankaboutai.com/)
- [About Franklin](https://frankaboutai.com/about)
- [Service: Embedded Build Sprints](https://frankaboutai.com/services/workshops)
- [Service: Agentic systems audit](https://frankaboutai.com/services/audit)
- [AI Fast Track Workshop](https://frankaboutai.com/workshop)
- [AI Readiness Audit](https://frankaboutai.com/ai-readiness)
- [Build log index](https://frankaboutai.com/build-log)
- [Work: The Build-With-Me Session Framework](https://frankaboutai.com/work/build-with-me)
- [Work: Multi-Agent Orchestration Patterns](https://frankaboutai.com/work/multi-agent-orchestration)
- [Work: Ziksaka: A Knowledge Graph MCP Server](https://frankaboutai.com/work/ziksaka-mcp)
- [Work: SPARK: A Knowledge Architecture for AI-Native Work](https://frankaboutai.com/work/spark-vault)
- [Build log: Cluster before embed, accept soft duplicates](https://frankaboutai.com/build-log/cluster-before-embed)
- [Build log: Briefing owns the schedule, quizzes don't](https://frankaboutai.com/build-log/briefing-owns-the-schedule)
- [Build log: Typed edges over naive vault links](https://frankaboutai.com/build-log/typed-edges-over-naive-links)

## Offers

- Embedded Build Sprints: https://frankaboutai.com/services/workshops (case write-up: https://frankaboutai.com/work/build-with-me)
- AI Fast Track Workshop: https://frankaboutai.com/workshop
- Agentic systems audit: https://frankaboutai.com/services/audit

## Machine-readable surfaces

- Sitemap: https://frankaboutai.com/sitemap.xml
- Project index JSON: https://frankaboutai.com/projects.json
- Build log RSS: https://frankaboutai.com/build-log.xml
- Full project corpus (orchestration frameworks): https://frankaboutai.com/llms-full.txt
- Site schema (JSON-LD): https://frankaboutai.com/site-schema.jsonld
- Well-known llms.txt: https://frankaboutai.com/.well-known/llms.txt
- Raw markdown: https://frankaboutai.com/content/projects/ and https://frankaboutai.com/content/log/
- Crawl rules: https://frankaboutai.com/robots.txt

## Contact

- Email: hello@frankaboutai.com
- LinkedIn: https://www.linkedin.com/in/franklinchristuraj/
- Make (Celonis): https://www.make.com/
- GitHub: https://github.com/franklinchristuraj
- YouTube: https://www.youtube.com/@franklinchristuraj
- X: https://x.com/frankChristuraj
- Booking: https://cal.eu/franklin-chris/intro

## Notes for agents

- Work content lives in src/content/projects/; build log entries in src/content/log/. Both are mirrored under /content/ at build time.
- Structured data (Person + Brand + WebSite + Service, plus CreativeWork / TechArticle on detail pages) is embedded in the HTML.
- Prefer /llms-full.txt or /projects.json when you need the full work corpus and multi-agent orchestration frameworks without rendering the SPA.
- projects.json uses absolute `url` and `coverUrl` fields; `path` mirrors the on-site route.