Person peeking from behind laptop
Open to Tampa hybrid/onsite and US remote roles

JC Gurdian

AI Engineer at yorCMO. Co-Founder and CTO of Gynka. Tampa, FL.

I design, ship, and run production LLM agent systems. At yorCMO I built 4 of 11 production agents on a platform serving 24 client organizations and led an outreach agent whose drafts are approved without edits about 90% of the time.

See the WorkEmail Me

Production Systems

Featured Projects

Production systems I built and run, plus open-source work you can read.

LinkyyorCMO • 202501 / 07

Linky: Autonomous LinkedIn Outreach Platform

  • Problem: fractional CMOs needed personalized LinkedIn outreach at a volume no one could write by hand.
  • Built: LLM agents that source about 6,000 companies and 12,000 prospects per client, research each prospect, and draft messages in the user's voice, with human approval before anything sends. I led the build and directed one engineer.
  • Result: about 90% of drafts approved without edits. Stack: LangGraph, Unipile, Supabase, Modal, per-agent cost tracking.
yorCMO
COREyorCMO • 202502 / 07

CORE Growth Platform + MCP Server

  • Built 4 of 11 production LLM agents serving 24 client organizations through 25+ integrations (HubSpot, WordPress, Google Analytics and Ads, Mailchimp, Klaviyo, Shopify, Notion, Microsoft Graph, and more).
  • Built and maintain the MCP server that exposes CORE's agents as tools inside Claude, plus the shared orchestration and tool-calling layer.
  • Guardrails, prompt-regression and A/B evals, RBAC with audit logs, and per-agent cost and latency controls.
yorCMO
PanelCastyorCMO • 202503 / 07

PanelCast: Meeting-to-Content Pipeline

  • Turns one recorded meeting into an on-brand blog post with images, a LinkedIn post for each attendee, and short-form video clips.
  • Builds a brand and voice profile for each person from their social posts and writing samples, so the output sounds like them.
yorCMO
GynkaGynka • 202504 / 07

Gynka: AI Fitness Coach

  • A performance coach for self-coached athletes: it reads wearable biomarkers and rewrites the day's session toward the user's goal.
  • I'm the only engineer: multi-agent engine on AWS (FastAPI, Pydantic AI, Supabase, ARQ/Redis), React Native iOS and Android app, RevenueCat billing.
  • Status: closed beta on TestFlight.
View Project
WingmanOpen Source • 202505 / 07

Wingman: MCP Bridge for Coding Agents

  • TypeScript MCP server that lets a remote AI assistant monitor and drive local Claude Code and Codex sessions.
  • Bearer-token auth, tunnel setup docs, published to npm as wingman-mcp.
View Project
Vault-AIOpen Source • 202506 / 07

Vault-AI: Local-First AI Platform

  • Runs local models on your own hardware. An LLM router sorts each query into one of seven task types and picks a model in under 500 ms, with regex fallback, caching, and automatic upgrades when a prompt overflows the context window.
  • Next.js and FastAPI, Ollama, ChromaDB RAG, SearXNG search, ComfyUI images. One-command Docker Compose deploy.
View Project
Bug ButlerOpen Source • 202507 / 07

Bug Butler: Slack to GitHub Issues

  • Slack bot that turns a plain-language bug report into a structured GitHub Issue. It asks follow-up questions when details are missing and shows a preview before filing.
  • Python, FastAPI, Slack Bolt, LiteLLM (works with multiple model providers), Supabase, with CI.
View Project
Python
TypeScript
LLMs
Agentic Workflows
RAG
MCP
LangGraph
LangChain
Pydantic AI
FastAPI
React
Next.js
React Native
AWS
Modal
Vercel
Docker
PostgreSQL
Supabase
Redis
GitHub Actions
Python
TypeScript
LLMs
Agentic Workflows
RAG
MCP
LangGraph
LangChain
Pydantic AI
FastAPI
React
Next.js
React Native
AWS
Modal
Vercel
Docker
PostgreSQL
Supabase
Redis
GitHub Actions

Career

Experience

Building production AI systems for real clients.

AI Engineer
AI Engineer
yorCMO | Sep 2025 to Present | Tampa, FL (remote)
  • ▸Led the build of Linky, an autonomous LinkedIn outreach platform on LLM agents. About 90% of its drafts are approved without edits. Directed one engineer.
  • ▸Built 4 of 11 production LLM agents on CORE, a multi-tenant marketing platform serving 24 client organizations through 25+ integrations.
  • ▸Built and maintain the MCP server that exposes CORE's agents as tools inside Claude, with guardrails, evals, RBAC, and per-agent cost controls.
Co-Founder and CTO
Co-Founder and CTO
Gynka | Jan 2025 to Present | Remote
  • ▸Only engineer on an AI fitness coaching app in closed beta (gynka.app).
  • ▸Designed and built the multi-agent coach engine on AWS: FastAPI, Pydantic AI, Supabase, ARQ/Redis, RevenueCat.
  • ▸Built the iOS and Android app in React Native and ship it through TestFlight to advisors and early users.
  • ▸The coach reads biomarker data from connected wearables and adjusts each workout toward the user's goal.
  • ▸Own schemas, APIs, infrastructure, and the release process.
Machine Learning Intern
Machine Learning Intern
AtomChat | Jul to Oct 2024 | Remote
  • ▸Built and deployed a customer-facing LLM agent for automotive dealerships (OpenAI API, LangChain, Pinecone, MySQL, Google Calendar) on WhatsApp, Instagram, Messenger, and Facebook.
  • ▸Built Scrapy ingestion pipelines and tuned vector queries to keep retrieval real-time.
Digital Transformation Intern
Digital Transformation Intern
Cargill | Jun to Aug 2023
  • ▸Built Power Apps workflow automations and Power BI adoption and ROI dashboards with international managers, reducing operational workload by 25%.

Introduction

About Me

Based in Tampa, FL. TCU B.S. Computer Information Technology, May 2025.

About Me

AI Engineer based in Tampa, FL

I'm an AI engineer in Tampa, FL. I joined yorCMO in September 2025, a few months after finishing my B.S. in Computer Information Technology at TCU, and I've spent the year building production LLM agent systems that fractional CMOs and professional firms use every day. I led the build of Linky, an autonomous LinkedIn outreach platform, built 4 of 11 production agents on CORE, a multi-tenant marketing platform serving 24 client organizations, and maintain the MCP server that exposes CORE's agents as tools inside Claude.

I'm also cofounder and CTO of Gynka and its only engineer: the React Native app, the multi-agent coaching backend, the database, and the AWS infrastructure.

I use Claude Code and Codex to move faster. Architecture, evals, guardrails, and what reaches production are my responsibility.

Languages: English and Spanish

JC Gurdian

Education & Certs

Credentials and training

  • ▸Texas Christian University, B.S. Computer Information Technology, May 2025
  • ▸Claude Code certification (Anthropic), earned through yorCMO's Anthropic partnership
  • ▸AWS Certified Cloud Practitioner

My Toolbox

What I build production systems with

AILLM agents, RAG, evals, guardrails, MCP, LangGraph, Pydantic AI, Anthropic and OpenAI APIs
BackendPython, FastAPI, TypeScript, Node.js, PostgreSQL, Supabase, Redis
FrontendReact, Next.js, React Native
CloudAWS, Modal, Vercel, Docker, GitHub Actions

Beyond the Code

Explore my interests and hobbies beyond the tech world.

Motocross🏍
Fitness🏋🏼
3D Printing🤖
Reading📚
Map of the Tampa Bay area
smiling memoji

Contact

Get in Touch

Open to Tampa hybrid/onsite and US remote AI engineering roles.

Send me a message

Contact Information

Social Media

LinkedInGitHub

Availability

Open to opportunities

Based in Tampa, FL. Open to Tampa hybrid/onsite and US remote AI engineering roles.