About Motion Lab
We are building the physical-world data acquisition layer for embodied AI in Latin America.
Motion Lab creates mobile infrastructure to collect localized, first-person video data from real Brazilian environments — helping robotics, AI, and multimodal model companies understand the physical world beyond generic datasets.
Core Thesis
Trampo is the acquisition engine.
The core asset is proprietary physical-world data from Latin America.
Motion Lab is not positioned as another gig app or SaaS tool. Trampo enables a distributed mobile collection network; Motion Lab builds the data infrastructure behind it.
Our goal is to help robotics, AI, and multimodal model companies access localized, consent-based, first-person data from real Brazilian environments.
Gabriel Gil
Founder & CEO
Founder-led infrastructure
A Brazilian data layer for physical AI.
Motion Lab was founded by Gabriel Gil to solve a specific bottleneck in embodied AI: robots and multimodal systems need high-quality, localized data from the real physical world. Gabriel’s background combines product testing, digital marketing, technical education, and startup execution — a practical mix for building a data acquisition layer from Brazil to the global AI and robotics market.
Our approach combines a mobile collection app, task design, contributor screening, quality validation, and dataset delivery into one acquisition pipeline.
POV
First-person data
BR
Localized context
B2B
AI & robotics
Founder background
Gabriel brings a mix of product testing, digital marketing, technical education, and startup execution to Motion Lab’s physical-world data mission.
Founder-led
CEO of Motion Lab since 2024
Product QA
AAA game playtesting and product feedback
Digital Marketing
AAS in Social Media Marketing • 3.96 GPA
Technical Foundation
Computer technician, web design, and game development
What we build
Infrastructure for localized physical-world datasets.
Motion Lab is developing the operational layer required to turn everyday smartphone users into structured, consent-based POV data collectors for AI and robotics companies.
Mobile acquisition
Trampo guides selected contributors through structured recording tasks using smartphones.
Quality validation
Data is designed to pass through instruction checks, quality rules, and approval workflows before delivery.
Embodied AI use cases
Datasets may support robotics, multimodal models, physical AI, simulation, and real-world behavior understanding.
Operating principles
Built for trust, consent, and useful data.
Consent-first
Participants must understand what they are recording and how approved data may be used.
Brazilian context
The environments, objects, homes, routines, and motion patterns reflect local reality.
Task structure
Collection is organized around specific physical tasks, instructions, and acceptance criteria.
B2B delivery
The goal is to create useful datasets for companies building robotics and AI systems.
For AI companies, robotics teams, and investors
Let’s talk about physical-world data from Brazil.
If you are building embodied AI, robotics systems, multimodal models, or data infrastructure, reach out directly to discuss partnerships, pilots, or Motion Lab’s roadmap.
Brazil is an underrepresented physical-world data market.
Motion Lab is building the infrastructure to make it accessible, structured, and useful for the next generation of AI systems.