Chris Sean Dabatos
I break down AI tools for developers who actually ship code.
Las Vegas, Nevada, United States
Actions
Chris is a Staff Developer Relations Engineer with a focus on AI & data infrastructure. He spends a lot of time building AI agents that he actually uses and speaking at meetups and conferences. He’s big on showing, not just telling. So expect real world examples and maybe a few stories about what didn’t work. Outside of work, he runs a YouTube channel, helps organize hackathons, and lives in Las Vegas.
Links
Area of Expertise
Topics
AI Memory Is Broken And How I Fixed It
Every AI app claims it can “remember” you. Most of them don’t.
When I built my voice AI agent, I watched it forget my salary, hallucinate my daughter’s name, and confidently describe a restaurant I never visited. So I stopped trusting the usual memory pattern: summarize the chat, store embeddings, and hope retrieval works later.
In this talk, I’ll walk through the three-layer memory architecture I built after those failures: a deterministic profile, a facts table, and vector search over transcript chunks. I’ll show where the first version broke, why whole-transcript embeddings made recall worse, and why the model should propose memory but never own truth.
You’ll leave with a practical way to design AI memory that survives corrections, voice latency, long-term context, and the boring facts users expect your app to get right.
Agents Need Computers. Humans Need Monitors.
I once told an agent to move a directory. It deleted the original first, then tried to copy it. The code was gone. What I needed was not a better answer. I needed a way back.
MCP is solving reach: a standard way for agents to touch files, services, databases, and private data. But reach is not safety. Every tool call MCP makes easier leaves consequences inside a real environment, and most agent tooling optimizes for capability, not trust.
Persistent agent work needs more than checkpoint and restore. A checkpoint without context is a restore button you are scared to press. The missing layer is restore confidence: what changed, whether verification passed, whether the app was healthy, and what restore will overwrite.
Through a live recovery demo of a real app in a persistent agent environment, this talk shows what that context layer looks like, why environment history matters beyond git, and why the MCP ecosystem needs monitors as much as it needs servers. A checkpoint with context is manufactured hesitation: the pause an agent will never take, built into the system around it.
Chris Sean Dabatos
I break down AI tools for developers who actually ship code.
Las Vegas, Nevada, United States
Links
Actions
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top