# Muratcan Koylan **Member of Technical Staff** at [Sully.ai](https://sully.ai) Toronto, ON, Canada - Website: [muratcankoylan.com](https://muratcankoylan.com) - X: [@koylanai](https://x.com/koylanai) - GitHub: [muratcankoylan](https://github.com/muratcankoylan) - LinkedIn: [muratcan-koylan](https://linkedin.com/in/muratcan-koylan) - Hugging Face: [MuratcanKoylan](https://huggingface.co/MuratcanKoylan) --- ## Summary Member of Technical Staff at Sully.ai, working on context engineering for clinical AI. Previously Entrepreneur in Residence at Antler, and before that AI Agent Systems Manager at 99Ravens AI, where I built the multi-agent platform, persona layer, agent skills library, and prompt defense stack. Seven years in B2B marketing before AI; Communication Design (HCI) background. --- ## About I'm a Member of Technical Staff at Sully.ai, a healthcare AI company building autonomous agents for hospitals. My work is on context engineering for clinical AI: AI Scribe and agent harnesses across medical specialties. Before Sully, I was Entrepreneur in Residence at Antler (Winter '26 Canada Residency, Jan–Feb 2026), working on auto-evolving context engineering skills for coding agents. I had several LOIs but withdrew to join Sully. Before Antler, I was AI Agent Systems Manager at 99Ravens AI in Toronto. I joined as a Prompt Designer (their first AI hire) and moved into the systems role about three months in. I built the multi-agent platform, the persona layer, the agent skills library, and the prompt defense stack after a meta-discussion jailbreak surfaced. Before that, seven years in B2B marketing: demand generation at Insider, lead operations and field marketing at Tmob, and communications plus internal AI work at Jonas Software. The path from marketer to builder ran through automating my own workflows with LLMs. What I think about: context engineering, multi-agent orchestration, persona embodiment, and how to make models behave like specific experts instead of generic assistants. --- ## Experience ### Sully.ai — Member of Technical Staff (2026 – Present) Toronto, ON (Remote) - Context engineering for clinical AI: AI Scribe and agent harnesses across medical specialties. - Work spans context architecture, dynamic prompting, prompt evaluation, and multi-agent systems. ### Antler — Entrepreneur in Residence (Jan 2026 – Feb 2026) Toronto, ON (On-site) - Accepted into the Antler Winter '26 Canada Residency with an auto-evolving context engineering skills MVP for coding agents. - Several LOIs; withdrew to join Sully. ### 99Ravens AI — AI Agent Systems Manager (2024 – 2026) Toronto, ON (Hybrid) - Architected the core multi-agent platform on LangGraph/LangChain, with end-to-end observability via LangSmith. - Built the persona layer and an internal agent skills library; codified expert heuristics and evaluation rubrics for predictable, auditable outputs. - Designed a five-layer prompt defense (instruction precedence, meta-request gating, hard deny-list, tool protocols, rationale policies) after a meta-discussion jailbreak surfaced. - Implemented context orchestration (file gating, dynamic prompt composition) to reduce context drift and improve response fidelity. - Shipped evaluation frameworks (Brief Analysis rubrics) and standardized strategic quality scoring across agents. - Led red teaming and remediation; wrote internal guidance on defense-in-depth for agent prompts and tools. ### Jonas Software — Marketing Communications & AI Transformation (2022 – 2024) - Built AI agents and generative workflows for marketing operations; led HubSpot onboarding. - Grew organic impressions 6x and lead generation 50% through AI-enabled content and automation. - Delivered executive AI training; established prompt standards for non-technical teams. ### Tmob (Thinks Mobility) — Senior Marketing Specialist (2021 – 2022) - Scaled lead generation 40% YoY with ABM and automation. - Grew LinkedIn 3x; owned martech stack, events, and product marketing campaigns. ### Insider — Marketing Associate (2020 – 2021) - Ran global demand-gen campaigns across 25 regions for Insider's AI Growth Management Platform. - Built automation, launched a podcast, and executed end-to-end B2B campaigns. --- ## Open Source & Projects ### Agent Skills for Context Engineering [GitHub](https://github.com/muratcankoylan/Agent-Skills-for-Context-Engineering) Transferable patterns for production AI agent systems on Claude Code, Cursor, or custom stacks. Covers context fundamentals, multi-agent patterns, memory, tool design, evaluation, and cognitive models. ### Ralph Wiggum Copywriter [GitHub](https://github.com/muratcankoylan/ralph-wiggum-marketer) Claude Code plugin: autonomous multi-agent copywriting loop. Learns voice, self-critiques, and rewrites until acceptance criteria pass. Memory persists through git commits and text files, not state. ### AI Investigator [GitHub](https://github.com/muratcankoylan/AI-Investigator) Autonomous research pipeline (Firecrawl + Claude) for enterprise AI case studies. Built in a weekend in response to an Arvid Kahl tweet. ### ReadWren (Nov 2025) [GitHub](https://github.com/muratcankoylan/ReadWren) Adaptive multi-agent interview that extracts literary preferences into structured profiles. LangGraph + Kimi K2; Redis checkpointing; rubric-driven generation. ### ActualCode (Sep 2025) [GitHub](https://github.com/muratcankoylan/actual_code) — 2nd place, Google Cloud × AI Tinkerers Hackathon Multi-agent system that analyzes GitHub repos and generates repo-specific coding assessments using Google's A2A protocol. ### Feed2Context (Sep 2025) [GitHub](https://github.com/muratcankoylan/feed2context) Chrome extension and local pipelines that convert LinkedIn/X posts into structured research reports. Groq Compound + Kimi K2 + Browser Use. ### Butterpath (Nov 2025) [GitHub](https://github.com/muratcankoylan/Butterpath) Prototype food tour planner using LangChain DeepAgents, Google Maps API, and Tavily research. ### Memory Routing Model (Nov 2025) [HuggingFace](https://huggingface.co/MuratcanKoylan/Marketing-Memory-Routing-8B) · [GitHub](https://github.com/muratcankoylan/memory-routing-agent) Fine-tuned an 8B Llama model on memory routing for marketing conversations using prompt distillation and reinforcement learning on Tinker. ### JeezAI (2023, paused) [GitHub](https://github.com/JeezAI/DSPy_matchmaking) AI career platform connecting students to AI internships via autonomous agents. DSPy matchmaking, multi-agent scraping. Paused to join 99Ravens. --- ## Skills - **Languages**: Python, TypeScript/JavaScript, SQL - **Frameworks**: LangGraph, LangChain, DSPy, FastAPI, Astro - **Agents**: Multi-agent orchestration, evaluation frameworks, context engineering, agent skills libraries, A2A protocol, DeepAgents - **Infrastructure**: Redis, Weaviate, PostgreSQL, Google Vertex AI, LangSmith - **Tooling**: OpenAI, Anthropic, Groq, Moonshot Kimi K2, Tavily, Firecrawl, Browser Use, Make, Zapier, Clay - **Data**: RAG, hybrid search (BM25 + vector), document gating - **Security**: Red teaming, instruction hierarchy, meta-request gating, prompt hardening --- ## Education **Ozyegin University** — Bachelor's, Communication Design (Human-Computer Interaction), 2014 – 2020 Full scholarship; student union; work shown in international exhibitions. --- ## Personal Born in Turkey. Read papers, build, play poker and FIFA, watch the Blue Jays. Some street photography on [VSCO](https://vsco.co/mkoylan/gallery). --- *Last updated: May 2026* *For work or collaboration, DM on [X](https://x.com/koylanai).*