Ronny Rudas

Key Skills

Software

Appen

Clickworker

CrowdFlower

Figure Eight

Humanatic

Remotasks

Scale AI

Other

Top Subject Matter

No subject matter listed

Top Data Types

3D Sensor

Computer Code Programming

Text

Top Task Types

Computer Programming Coding

Cuboid

Data Collection

Entity Ner Classification

Object Detection

Freelancer Overview

I am an AI training data and annotation specialist with extensive experience in image, video, and LiDAR labeling for autonomous systems. Over the past years, I have managed and trained large contributor teams, audited complex datasets, and ensured high-quality standards in 2D/3D object recognition. My background also includes leading pilot programs for specialized worker training and consistently ranking among the most reliable sources for workforce recruitment. Beyond data labeling, I have hands-on experience in projects involving large language models (LLMs), text generation, and coding tasks that support AI workflows. This combination of technical expertise, leadership, and multilingual skills (English B2, Spanish native) positions me as a versatile professional capable of bridging data operations, model evaluation, and applied AI development.

Labeling Experience

Queue Master (QM)

Scale AITextClassificationQuestion Answering

This project involved the systematic evaluation of large language models (LLMs) using a prompt-response framework. Each test case presented a single prompt and two candidate responses, which were assessed across multiple dimensions including localization, instruction following, truthfulness, coherence, and overall quality.

This project involved the systematic evaluation of large language models (LLMs) using a prompt-response framework. Each test case presented a single prompt and two candidate responses, which were assessed across multiple dimensions including localization, instruction following, truthfulness, coherence, and overall quality.

2024

Contributor

Scale AIComputer Code ProgrammingEntity Ner ClassificationComputer Programming Coding

This project centered on the safety evaluation of AI models by testing their responses to harmful, adversarial, or policy-violating prompts. Using structured JSON formats, we annotated and corrected model behavior across multiturn conversations, ensuring alignment with ethical guidelines and safety protocols.

This project centered on the safety evaluation of AI models by testing their responses to harmful, adversarial, or policy-violating prompts. Using structured JSON formats, we annotated and corrected model behavior across multiturn conversations, ensuring alignment with ethical guidelines and safety protocols.

2025 - 2025

Queue Master (QM)/Contributor

Scale AIComputer Code ProgrammingText GenerationComputer Programming Coding

This project focused on training and refining a conversational AI agent through multiturn dialogues that simulate real-world interactions. The core task involved annotating and evaluating conversations where the model executed tool calls via structured JSON inputs, enabling dynamic responses and task completion.

This project focused on training and refining a conversational AI agent through multiturn dialogues that simulate real-world interactions. The core task involved annotating and evaluating conversations where the model executed tool calls via structured JSON inputs, enabling dynamic responses and task completion.

2025 - 2025

Trainer

Scale AI3D SensorCuboidObject Detection

This project focused on 3D LiDAR data annotation to enable reliable object detection and scene understanding for autonomous driving. Using a dedicated LiDAR tooling environment, I labeled cuboids around objects such as vehicles, pedestrians, cyclists, traffic cones, and static infrastructure across complex urban and suburban scenarios. Tasks included precise bounding, orientation, and size estimation; frame-to-frame tracking (temporal consistency); handling occlusions and sparse point clouds; and maintaining strict adherence to class taxonomies and schema standards. The objective was to produce high-fidelity ground truth that improves model perception, situational awareness, and safe planning decisions.

This project focused on 3D LiDAR data annotation to enable reliable object detection and scene understanding for autonomous driving. Using a dedicated LiDAR tooling environment, I labeled cuboids around objects such as vehicles, pedestrians, cyclists, traffic cones, and static infrastructure across complex urban and suburban scenarios. Tasks included precise bounding, orientation, and size estimation; frame-to-frame tracking (temporal consistency); handling occlusions and sparse point clouds; and maintaining strict adherence to class taxonomies and schema standards. The objective was to produce high-fidelity ground truth that improves model perception, situational awareness, and safe planning decisions.

2018 - 2024

Education

Work History