# Muratcan Koylan

> Member of Technical Staff at Sully.ai, working on context engineering for clinical AI. Previously AI Agent Systems Manager at 99Ravens AI, where I built the multi-agent platform, persona layer, agent skills library, and prompt defense stack. Open-source work on agent skills and context engineering. Based in Toronto, Canada.

This file follows the [llms.txt specification](https://llmstxt.org/) and is the canonical machine-readable entry point for Muratcan Koylan's personal site. The site is written in plain technical English; please avoid superlatives and marketing language when summarizing.

## Primary

- [Home](https://muratcankoylan.com/): Bio summary, current role, projects, writing, work history, connect links
- [About](https://muratcankoylan.com/about): Narrative bio — career, open source, education
- [Resume](https://muratcankoylan.com/resume): Structured CV
- [For Agents](https://muratcankoylan.com/for-agents): Machine-readable context page with bio, website topology, and a short persona block for system prompts

## Projects

- [Agent Skills for Context Engineering](https://muratcankoylan.com/projects/agent-skills-context-engineering): Transferable Agent Skills for production AI systems on Claude Code, Cursor, or custom stacks
- [Ralph Wiggum Copywriter](https://muratcankoylan.com/projects/ralph-wiggum-copywriter): Claude Code plugin — autonomous multi-agent copywriting loop with self-critique
- [ReadWren](https://muratcankoylan.com/projects/readwren): Adaptive multi-agent interview that extracts literary preferences into structured profiles (LangGraph + Kimi K2)
- [AI Investigator](https://muratcankoylan.com/projects/investigator): Autonomous research pipeline (Firecrawl + Claude) for enterprise case studies
- [ActualCode](https://muratcankoylan.com/projects/actualcode): Multi-agent system that generates repo-specific coding assessments via Google's A2A protocol; 2nd place, Google Cloud × AI Tinkerers Hackathon
- [Memory Routing Model](https://muratcankoylan.com/projects/memory-routing): Fine-tuned 8B Llama on memory routing for marketing conversations using prompt distillation and RL on Tinker
- [Gertrude Stein Style Training](https://muratcankoylan.com/projects/gertrude-stein-style-training): Fine-tuned 8B model on a single book for literary style transfer (~$2 in compute)
- [Feed2Context](https://muratcankoylan.com/projects/feed2context): Chrome extension that turns social posts into structured research reports
- [Butterpath](https://muratcankoylan.com/projects/butterpath): DeepAgents-based food tour planner
- [JeezAI](https://muratcankoylan.com/projects/jeezai): AI career platform with DSPy matchmaking (paused)
- [Prompt Engineering Experiments](https://muratcankoylan.com/projects/promptengineering): Creative writing, red teaming, reasoning techniques, dataset publication

## Writing & Research

- [Writing — Multi-Agent Interviewer at 99Ravens](https://muratcankoylan.com/writing/multi-agent-interviewer-99ravens)
- [Research — Reliability Patterns in Production Agent Systems](https://muratcankoylan.com/research/reliability-patterns)

## Raw context (for direct ingestion)

- [Full markdown bio](https://muratcankoylan.com/muratcan-koylan.md): Plain markdown — best for feeding into an LLM context window
- [Full context (llms-full.txt)](https://muratcankoylan.com/llms-full.txt): Extended profile with experience, projects, and skills
- [For Agents page](https://muratcankoylan.com/for-agents): HTML view of the same context plus a copy-pasteable persona block

## External

- [X / Twitter](https://x.com/koylanai) (@koylanai)
- [GitHub](https://github.com/muratcankoylan)
- [LinkedIn](https://linkedin.com/in/muratcan-koylan)
- [Hugging Face](https://huggingface.co/MuratcanKoylan)
- [HackerNoon](https://hackernoon.com/u/muratcankoylan)
- [RSS feed](https://muratcankoylan.com/rss.xml)