Mishka Alditya Priatna

Key Skills

Software

CVAT

LabelImg

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Image

Video

Top Task Types

Bounding Box

Object Detection

Tracking

Freelancer Overview

I have extensive experience in data labeling and AI training data preparation, particularly for object and anomaly detection tasks. At AI Research Center ITB – MarkAny, I annotated and analyzed large-scale video datasets for crime detection, ensuring accuracy, consistency, and contextual relevance of metadata and labels. I also worked as a freelance Data Analyst for a Swiss-based consulting firm, where I conducted video data annotation for industrial pipeline inspections, timestamped key anomalies, and collaborated with global teams to refine supervised learning datasets. My strength lies in producing high-quality, structured training data that enables reliable AI model performance. I bring strong attention to detail, critical thinking, and validation skills, which ensure that datasets not only meet technical requirements but also align with domain-specific needs. This combination of technical expertise and practical annotation experience positions me to deliver training data pipelines that directly improve the accuracy and robustness of AI systems.

Labeling Experience

Crime Detection

CVATVideoBounding Box

At MarkAny – AI Research Center ITB, I contributed to a large-scale project focused on annotating video datasets for AI-based crime detection. The project scope involved labeling multiple object categories such as humans, vehicles, and contextual environmental cues, with the goal of creating high-quality supervised learning data. The project size encompassed more than 300 hundreds of video frames across diverse real-world crime scenarios, requiring systematic annotation strategies and collaboration with a research team. My specific tasks included applying bounding boxes, classification tags, and metadata enrichment, while handling challenging cases such as occlusions and low-light conditions. To ensure reliability, I adhered to strict quality measures such as cross-validating annotations, keeping bounding box as tight as possible, and implementing feedback-driven refinements to be more accurate, consistent, and contextually relevant datasets that directly supported AI model training.

At MarkAny – AI Research Center ITB, I contributed to a large-scale project focused on annotating video datasets for AI-based crime detection. The project scope involved labeling multiple object categories such as humans, vehicles, and contextual environmental cues, with the goal of creating high-quality supervised learning data. The project size encompassed more than 300 hundreds of video frames across diverse real-world crime scenarios, requiring systematic annotation strategies and collaboration with a research team. My specific tasks included applying bounding boxes, classification tags, and metadata enrichment, while handling challenging cases such as occlusions and low-light conditions. To ensure reliability, I adhered to strict quality measures such as cross-validating annotations, keeping bounding box as tight as possible, and implementing feedback-driven refinements to be more accurate, consistent, and contextually relevant datasets that directly supported AI model training.

2024 - 2025

Pipeline Inspection

Internal Proprietary ToolingVideoPoint Key PointObject Detection

I worked on an international project annotating video datasets for AI-based defect detection in industrial pipeline inspections. The project scope focused on labeling anomalies within long video streams to support supervised learning models for infrastructure monitoring. The project size involved hours of inspection footage and more than 700 of timestamped anomalies across multiple defect categories, requiring careful attention to domain-specific details. My specific tasks included segmenting video sequences, tagging defect types, and providing precise timestamps to highlight irregularities, ensuring the dataset captured both spatial and temporal features. To guarantee quality, I have to do keep learning on specific domain knowledge, conduct structured evaluation methods, perform consistency checks, and collaborate with global teams in iterative feedback loops, resulting in high-integrity datasets that improved model performance in detecting industrial defects.

I worked on an international project annotating video datasets for AI-based defect detection in industrial pipeline inspections. The project scope focused on labeling anomalies within long video streams to support supervised learning models for infrastructure monitoring. The project size involved hours of inspection footage and more than 700 of timestamped anomalies across multiple defect categories, requiring careful attention to domain-specific details. My specific tasks included segmenting video sequences, tagging defect types, and providing precise timestamps to highlight irregularities, ensuring the dataset captured both spatial and temporal features. To guarantee quality, I have to do keep learning on specific domain knowledge, conduct structured evaluation methods, perform consistency checks, and collaborate with global teams in iterative feedback loops, resulting in high-integrity datasets that improved model performance in detecting industrial defects.

2023 - 2024

Education

Work History