Johann Cardenas

Key Skills

Software

AWS SageMaker

Dataloop

Google Cloud Vertex AI

Remotasks

Scale AI

Toloka

Top Subject Matter

No subject matter listed

Top Data Types

3D Sensor

Computer Code Programming

Image

Top Task Types

Computer Programming Coding

Evaluation Rating

Fine Tuning

Object Detection

Prompt Response Writing SFT

Freelancer Overview

I am an experienced software engineer and AI training contributor with over 10 years of professional development experience and a recent focus on AI-driven development automation and prompt engineering. At Outlier and Upwork, I have participated in diverse data labeling projects involving LLM evaluation, golden response creation, rubric-based grading, and multi-step reasoning tasks for code synthesis, front-end UI analysis, and scientific problem-solving. My work has spanned domains like physics simulations, debugging trajectories, and math understanding with multimodal input (e.g., graphs, handwritten notes).

Labeling Experience

Newton Buffet

Scale AIComputer Code ProgrammingComputer Programming Coding

This work involves designing high-difficulty programming prompts that challenge AI models to generate dynamic 2D and 3D physics simulations using technologies like p5.js, WebGL, or similar environments. Your responsibilities include:

This work involves designing high-difficulty programming prompts that challenge AI models to generate dynamic 2D and 3D physics simulations using technologies like p5.js, WebGL, or similar environments. Your responsibilities include:

2025 - 2025

Radiant Lampfish

Scale AIComputer Code ProgrammingPrompt Response Writing SFT

This project focuses on collaborative front-end development using a coding model (LLM). The main goal is to iteratively build and refine interactive, visually compelling web applications with the help of AI. The workflow involves multiple cycles of coding, reviewing, and enhancing output for quality and functionality.

This project focuses on collaborative front-end development using a coding model (LLM). The main goal is to iteratively build and refine interactive, visually compelling web applications with the help of AI. The workflow involves multiple cycles of coding, reviewing, and enhancing output for quality and functionality.

2025 - 2025

Nexar AI Annotation (IDOT Project)

DataloopImagePolygonPolyline

Annotated and classified roadway images based on snow coverage levels using the IDOT six-tier labeling system (Codes 1–6). This involved carefully reviewing dashcam-style images and accurately tagging conditions such as "All Clear," "Scattered Snow," "50% Bare," and "Snow Covered" based on visible lane coverage, wheel paths, and road markings. Each image required consistent attention to weather patterns, lighting variation, and surface texture to align with standardized snow classification criteria.

Annotated and classified roadway images based on snow coverage levels using the IDOT six-tier labeling system (Codes 1–6). This involved carefully reviewing dashcam-style images and accurately tagging conditions such as "All Clear," "Scattered Snow," "50% Bare," and "Snow Covered" based on visible lane coverage, wheel paths, and road markings. Each image required consistent attention to weather patterns, lighting variation, and surface texture to align with standardized snow classification criteria.

2025 - 2025

Multi-Class Visual Annotation for Street-Level Scene Understanding

Internal Proprietary ToolingImageBounding BoxPolygon

At Spare5, I contributed to high-precision data labeling tasks focused on improving visual perception systems used in AI and computer vision, particularly for autonomous vehicles and mapping technologies. My responsibilities included: Object annotation and classification of a wide range of entities in street-level imagery, including: Vehicles (cars, trucks, motorcycles) Street infrastructure (signs, lanes, lights) Pedestrians and cyclists Animals (in rural or urban environments) Miscellaneous objects (trash bins, mailboxes, road debris) Creating accurate bounding boxes, segmentations, and multi-label tags using proprietary or open-source annotation platforms Performing quality assurance reviews to validate consistency, label correctness, and adherence to detailed class taxonomies Participating in edge-case discovery by flagging ambiguous or novel visual situations (e.g., occluded signs, unusual vehicles, or animal-road interactions)

At Spare5, I contributed to high-precision data labeling tasks focused on improving visual perception systems used in AI and computer vision, particularly for autonomous vehicles and mapping technologies. My responsibilities included: Object annotation and classification of a wide range of entities in street-level imagery, including: Vehicles (cars, trucks, motorcycles) Street infrastructure (signs, lanes, lights) Pedestrians and cyclists Animals (in rural or urban environments) Miscellaneous objects (trash bins, mailboxes, road debris) Creating accurate bounding boxes, segmentations, and multi-label tags using proprietary or open-source annotation platforms Performing quality assurance reviews to validate consistency, label correctness, and adherence to detailed class taxonomies Participating in edge-case discovery by flagging ambiguous or novel visual situations (e.g., occluded signs, unusual vehicles, or animal-road interactions)

2017 - 2020

Education

Work History