Diego Pousa

Available to hire

I am Diego Pousa, a specialized data collection vendor focused on ego-centric video data sets for computer vision projects.

About

I am Diego Pousa, a specialized data collection vendor focused on ego-centric video data sets for computer vision projects. I lead a pre-vetted team of contributors and coordinate data collection across diverse environments to deliver high-quality, high-volume datasets.

I have a proven track record establishing internal QA pipelines to achieve 100% acceptance rates for AI training data. I implement secure, efficient workflows such as a single point of entry to protect accounts and data integrity, and I manage hardware logs, logistics, and payout structures for the local team to keep operations smooth and compliant.

  • Joined January 2026

Industry experience

  • Media & Entertainment
  • Software & Internet
  • Professional Services

More experts in Cordoba, Argentina

Projects · 4

High-Fidelity POV Sample: Outdoor Patio Maintenance & HDR Lighting

Dataset Sample: Raw ego-centric (POV) footage capturing maintenance tasks in an outdoor patio environment. Technical Complexity: High Contrast Lighting (HDR): Demonstrates sensor performance in direct sunlight with hard shadows (critical for outdoor computer vision). Fluid Dynamics: Interaction with water, foam, and cleaning tools in an uncontrolled environment. Environmental Elements: Exposure to wind and outdoor acoustics. Use Case: Ideal training data for Outdoor Service Robots, Environmental Perception Models, and autonomous cleaning systems. Production Capacity: My team operates in diverse semi-outdoor and outdoor residential perimeters (Gardens, Patios, Driveways). Outdoor NaturalLight HDR Robotics Cleaning video

  • Content Producer
High-Fidelity POV Sample: Delicate Object Handling & Sorting

Dataset Sample: Raw ego-centric (POV) footage capturing the disassembly and sorting of holiday decorations. Technical Complexity: Fragile Object Handling: Demonstrates careful manipulation of glass/delicate ornaments (Grip force estimation data). Deformable Linear Objects (DLOs): Untangling and winding light strings/wires—a critical benchmark task for robotic manipulation. Occlusion: High levels of object overlap (branches vs. ornaments) providing robust training data for segmentation models. Use Case: Highly relevant for Logistics AI, Warehouse Packing Robots, and Fine Motor Control training. Production Capacity: My team can replicate complex sorting and packing tasks across varying object types and environments. Robotics Logistics FineMotorSkills Sorting ComputerVision video

  • Content Producer
High-Fidelity POV Sample: Household Maintenance & Reflective Surfaces

Dataset Sample: Raw ego-centric (POV) footage capturing household maintenance tasks in a bathroom environment. Technical Complexity: Reflective Surfaces: Demonstrates clear data capture handling specular reflections (mirrors, chrome faucets, ceramic) which are critical edge-cases for Computer Vision. Deformable Object Interaction: Manipulation of cleaning cloths and fluids. Fine Motor Control: Scrubbing, wiping, and rinsing actions. Use Case: Essential training data for Service Robotics, Home Automation AI, and Activity Recognition models. Production Capacity: My team can scale this specific "Home Service" dataset to hundreds of hours, covering various bathroom layouts and cleaning protocols. Robotics HomeService Cleaning ComputerVision EdgeCaseData video

  • Content Producer
High-Fidelity POV Sample: Complex Hand-Object Interaction (Kitchen)

Dataset Sample: Raw ego-centric (POV) footage demonstrating fine motor skills and appliance interaction in a kitchen environment. Technical Specs: Resolution: High Definition (Raw export) Perspective: First-Person (Head-mounted stabilization) Scenario: Multi-step food preparation (Electric Juicer usage) Why this matters for AI: This footage is optimized for training Computer Vision models in Human-Object Interaction (HOI), Action Recognition, and Robotic Imitation Learning. It features clear occlusion handling (hands over objects) and stable framing essential for annotation. Production Capacity: I manage a team equipped to generate 50+ hours per week of similar task-based scenarios across diverse environments (Kitchen, Garage, Retail, Outdoor). ComputerVision video DataCollection POV MachineLearning VideoProduction

  • Content Creator
  • Content Producer

Work experience

Lead Data Collection Vendor · Confidential AI ProjectJan 2024 – Present · 2 yrs 11 mos

Managed a team of contributors collecting ego-centric video data for a major US-based AI Robotics project. Implemented a strict single point of entry system to maintain account security and data integrity. Achieved high-quality metrics by filtering out shaky, low-light, or irrelevant footage prior to submission. Managed hardware logs, logistics and payout structures for the local team.

Hire Diego Pousa today

Post a job and invite Diego, or contact them directly. No hidden commission.

Post a job