Felipe Hlibco — AI, Engineering Leadership & DevRel

/home

/posts

/projects

/reading

/skills

/tools

/talks

/about

~/ Felipe Hlibco | Tech Talks

Felipe Hlibco | Tech Talks

Claude Code Hack Day #

Presented “Architecting Multi-Agent Developer Workflows” to a sold-out audience of 260 engineers at the AWS Builder Loft. The talk walked through how to decompose complex developer tasks into orchestrated agent pipelines — covering tool selection, context management, long-term memory, exit criteria and the failure modes that show up when you let multiple LLM-based agents coordinate in production. Most of the Q&A focused on security boundaries between agents, which tracks with the broader industry concern around agentic systems operating without proper guardrails.

Conversational AI Agents #

Co-hosted this panel as part of the Conversational AI Mastermind series I help run in Palo Alto. The discussion centered on where conversational AI agents are actually being deployed versus where they’re still just demos — particularly in financial services and customer authentication scenarios. I shared our experience building RCS-based verification flows that replace SMS one-time passwords, which are trivially interceptable. Good discussion with people from the applied research side about how to make agent-to-agent handoffs auditable.

Conversational AI Agents #

The panel covered the emerging landscape of AI agents in messaging: what’s real, what’s vaporware, and what infrastructure is missing. I spoke about the gap between what LLMs can generate and what enterprises will actually trust in a customer-facing messaging channel, drawing on my years experience running the RCS for Business developer programs at Google and previously building the Google Business Messaging developer ecosystem.

DevRel-ativity #

Panel on the evolving role of Developer Relations Engineers in the age of AI. Hosted at Cloudflare’s SF office with other DevRel leaders from across the Bay Area. I talked about what I learned scaling Google’s RCS (and previously Google Business Messages) developer program across dozens of countries — specifically, how the playbook changes when you’re not just teaching developers to use an API, but convincing entire carrier ecosystems to adopt a new messaging standard. The conversation kept coming back to whether traditional DevRel is even the right model when AI tools are replacing tutorials, sample code and technical writing that used to be the backbone of the job.

CalibrateAI - Hack Day Micro-Conference #

Served as a judge for this AI-focused hackathon at the AWS Builder Loft. Teams had a single day to build working AI prototypes. I evaluated projects on technical depth, feasibility, and whether they solved a real problem or just wrapped a prompt around an API call. Several teams built agent-based systems, and the judging conversations kept circling back to reliability — how do you trust an AI agent to take actions on behalf of a user without proper validation layers? That question is basically my day job.

Rebuilding LA - Hack Night at GitHub #

Judged a hackathon organized in response to the LA wildfires, hosted at GitHub’s SF headquarters. The challenge was to build tools that could help with disaster relief and community rebuilding. I focused my evaluation on whether the solutions could actually be deployed — not just whether the demo looked good. A few teams built coordination platforms for volunteer efforts and resource allocation, which reminded me of the early days building Doare.org and the infrastructure challenges that come with mobilizing people at scale during a crisis.

AI Leaders Chat & Network at Cloudflare #

Panel discussion with AI leaders at Cloudflare’s office in San Francisco. The format was informal — short intros from each panelist, then open conversation with the audience. I spoke about the security gap in how AI agents interact with external systems, particularly through messaging channels that were never designed for machine-to-machine trust. The room had a mix of founders, engineers, and investors, and the most productive exchanges were around what “safe” actually means when an AI agent can send messages, make API calls, and take actions on behalf of a business.

Gogole I/O 2022 #

Spoke about bringing conversational AI to Google Search, Maps, and Shopping with the power of ML and Google AI. With Google’s Business Messages, customers can have engaging conversations throughout the Google ecosystem where they can shop online, make appointments, or get customer support. Demonstrated how a suite of tools empower developers to build connected and intelligent chat automations that allow customers to connect with brands wherever they are in their purchasing journey.

Conversational Interaction Conference #

  • Location: San Jose, CA
  • Date: April 12-13, 2022
  • Role: Speaker
  • Topic: Lucrative Conversational Commerce with New Chatbots
  • Website: Speech Technology
  • Hosted by: AVIOS
  • Attendees: ~250

Spoke at the Conversational Interaction Conference about the state of Google Business Messages adoption and what it means for the next generation of conversational interfaces. At the time, I was leading the Google Business Messaging developer ecosystem at Google, and this talk covered the technical and ecosystem challenges of getting enterprises aligned around a richers and more secure messaging channel. The audience was a mix of speech technology researchers, telecom engineers, and product people — a different crowd than the typical Silicon Valley meetup, and the questions reflected it. Lots of focus on interoperability, engagement metrics and security measures with agent identity.

←
Felipe Hlibco | Tools
Felipe Hlibco | Skills
→
back to top