AI product engineering
Thought leadership & speaking
REF 10.3157° N / 123.8854° E
UX design → product leadership → applied AI engineering
I’ve worked on every side of the AI product: designing the experience, leading the team, and engineering the system. My talks connect those perspectives to show where human judgment, trust, and taste matter most.
SUBJECT / JV-0101 / Point of view
When the machine drafts first, the work does not disappear. It moves. Taste, feedback, trust, and accountability become more important because every team can now produce before it has learned what deserves to exist.
I see that shift from four sides of the work: UX design, venture building, product leadership, and applied AI engineering. My talks connect them to show where AI-native products hold, where they break, and what we still owe the person on the other side of the interface.
02 / Signature talks
AI makes output cheap. Taste, evidence, and accountability become the scarce layer. A field report on how product, design, and engineering work changes when the first draft is no longer human.
A practical method for testing intent, behavior, and edge cases before polishing screens. Developed through AI product work and shared with the design community at Silicon Valley AI Nights.
The durable product is the system around the model: context, tools, permissions, feedback, verification, and interface. What teams should design when intelligence becomes a commodity.
Application, machine, and organization: the three levels builders must connect as agents move from a feature into real work. Based on the Builder’s Guide to Agentic Applications.
03 / Recently published
The Builder’s Guide
I recently wrote a practical CopilotKit × LangChain guide to the three surfaces builders must connect when creating, scaling, and operating intelligent agents.
Level 1Application agents
Level 2Machine-level agents
Level 3Organizational agents
Application, machine, and organization are not separate concerns. The same systems thinking behind the book informs every talk: practical patterns, clear boundaries, and a path from prototype to organizational capability.
04 / Work behind the ideas
My talks are field reports from building at the frontier: public code, product launches, agent infrastructure, interfaces, and founder bets—not observations from the sidelines.
Explore the full portfolio1.5K GitHub stars
An open-source framework for agents that generate rich, interactive interfaces—not just text responses.
Slack + Teams
Agent infrastructure and a public reference agent for the communication surfaces where work already happens.
Google Cloud Next 2026
A builder-facing environment that made an agent-to-interface protocol tangible and explorable.
Database core + cloud
A folder-based database and platform for humans and agents working across sessions.
First browser in one week
Infrastructure for testing, reviewing, coordinating, and governing the work AI agents produce.
05 / Why Jerel
I didn’t arrive at AI from one discipline. I followed the product from the interface, through the organization, and into the system underneath it.
That path lets me translate between designers, product leaders, founders, and engineers. Every talk pairs a sharp thesis with receipts: shipped products, real artifacts, operating decisions, and the failures that forced the lesson.
Started with the person on the other side of the interface: their context, decisions, friction, and trust.
Built across products at Full Scale Ventures, learning to move from ambiguity to evidence without losing the user.
Led large multidisciplinary teams at GoTeam and learned how product judgment scales through people and operating systems.
Now a Staff AI Engineer at CopilotKit and founder of Folderbase.ai, building the systems behind AI-native experiences.
23Documented talks
and workshops
10×Google Developer
Group speaker
4×San Francisco
hackathon winner
AuthorBuilder’s Guide to
Agentic Applications
06 / Formats
Every engagement is shaped around the event theme and audience—without turning the stage into a product pitch.
25–45 min
One sharp thesis, grounded in shipped systems and practical decisions, shaped around the event’s central question.
30–60 min
A candid conversation about what is changing, what remains unresolved, and what builders should do differently now.
Half or full day
A hands-on session where teams turn AI product ideas into testable conversations, behaviors, boundaries, and interface decisions.
07 / Speaking experience
23 engagementsInternational stages with audiences up to 1,000, prompt prototyping at Silicon Valley AI Nights, and monthly hands-on workshops with the AI Cebu community.
AIPC · Build with AI by Google
Executive AI enablement
Panelist
Cebu Technological University · 2024
GDG Bogor · Indonesia
Hands-on workshop
Google I/O Extended Cebu · 2024
Technical workshop
Google Mountain View
Google Developer Groups
Silicon Valley AI Nights · Frontier Tower
AI careers talk
Google Developer Groups
Build with AI
PizzaPy: LLMs
Applied AI product demo
Google Developer Groups
Leadership talk
Technical workshop
The Frame · 2024
Education talk
AI Pilipinas Cebu
GDG DevFest Cebu · 2024
08 / For organizers
The event theme, audience, format, and the question the room should leave debating.
One clear realization, a narrative built for the room, and no generic AI-is-changing-everything filler.
Shipped products, real artifacts, operating decisions, and honest failures—not a disguised product pitch.
Keynote, fireside, panel, or workshop with practical language and room for hard questions.
Available for remote and in-person engagements