2024 →
Emilia AI
Co-founder & CTO
AI copilots for restaurant frontline teams: cameras, POS data and humans in the loop. Real-time video, computer vision, VLMs and edge connectivity.
- computer vision
- VLM
- GStreamer
- edge
- Kubernetes
- AWS
For 15+ years I've built open-source real-time systems. In 2024 I co-founded Emilia AI, where the cameras, point-of-sale data and real-time signals restaurants already have turn into timely action for frontline teams, with humans in the loop.
I wrote the first commit and most of the platform: real-time video ingestion and recording, computer vision and vision-language model pipelines, secure edge connectivity to cameras behind NAT, and the cloud platform on Kubernetes and AWS. In 2026 Emilia raised a US$1.5M pre-seed round led by Prosus Ventures, and we're now a team of 10.
Before that I was open-source maintainer at Vocode (YC W23), founded WunderComm and ArtisanLabs, was CTO at RisePal, and spent 13 years as CTO of Sapian building open-source real-time communications. I organize the Supabase meetups in Medellín, and I've been part of the city's Linux and free-software community since the Asterisk days.
Projects
2026 →
pi-animations
Author
26 terminal animations for the Pi coding agent's thinking, working and tool indicators, from demoscene fire to Matrix rain, in ANSI true color. Install: pi install npm:pi-animations.
- AI agents
- TypeScript
- TUI
- open source
2026 →
pi-rewind
Author
Checkpoint and rewind for the Pi coding agent: roll files and conversation back to any earlier point.
- AI agents
- TypeScript
- open source
2026 →
bzz
Author
A human-first terminal client for Buzz: fast keyboard navigation, local offline history, native Nostr authentication.
- Rust
- TUI
- Nostr
2023–2024
Vocode (YC W23)
Open-source maintainer
Maintainer of the open-source voice-AI framework: Pydantic v2 support, OpenAI upgrades, CI, docs and a voice-RAG demo; community and events.
- voice AI
- LLM
- Python
2009–2022
Sapian
CTO
13 years building open-source real-time communications: cloud call center, corporate voice, WebRTC and SIP. Grew ARR by US$374K and cut latency 30% with a hybrid-cloud RTC stack.
- WebRTC
- SIP
- Asterisk
- Kubernetes
2022
whisper-large-v2-es
Author
Whisper large-v2 fine-tuned for Spanish at the Hugging Face × Lambda community event. WER 5.29 on Common Voice 11 (es).
- ASR
- Whisper
- Hugging Face
Pi extensions
❯ pi install npm:pi-rewindSix extensions for the Pi coding agent, on npm: undo (pi-rewind), 26 terminal animations, context health, worktree awareness and minimal web providers.Explore the extensions →Writing
De RisePal a ArtisanLabs: Innovación en la Era de la IA de Consumo2023 · ES
From RisePal to ArtisanLabs: Innovation in the Era of Consumer AI2023 · EN
Extending Workshop Learnings: From Supabase and PGvector to LangChain2023 · EN
Vocal Resonance2023 · EN
Archive
ArtisanLabs Tech Blog2023 · 5 · EN/ES
Arpagon Data Analysis2015–2016 · 17 · EN/ES
Medellín Asterisk2008–2009 · 17 · ES
Community
Organizer · LW14 Supabase Meetup Medellín (Supabase + v0 + MCP)2025-04
Organizer · LW13 Supabase Meetup Medellín (Elixir & Realtime)2024-12
Organizer · Supabase Meetup Medellín (Postgres, realtime, auth, storage)2024-08
Speaker · UbuConLA 2024: fully open-source conversational AI on Linux2024
Long-time member of Medellín's Linux & free-software community
Press
Emilia AI raises $1.5M led by Prosus VenturesProsus · 2026-09
Startup colombiana Emilia AI logra US$1,5 millones en ronda pre-seedThe Intelligent · 2026-09
Now
- Building Emilia AI: copilots for restaurant frontline teams.
- Maintaining pi-animations, pi-rewind and bzz.
- Rebuilding this site and bringing back 15 years of writing (Asterisk, data analysis, ArtisanLabs).
Updated