Wes

Agency

Kuala Lumpur, Malaysia

www.xtecq.com

Key Skills

Software

Appen

OneForma

CVAT

Top Subject Matter

No subject matter listed

Top Data Types

Image

Video

Text

Top Task Types

Data Collection

RLHF

Classification

Evaluation/Rating

Question Answering

Company Overview

XTECQ is a Southeast Asia-focused AI data operations company specializing in multilingual data annotation, RLHF, AI evaluation, and conversational AI datasets for modern machine learning systems. Our mission is to help AI models better understand Southeast Asian languages, cultures, and real-world communication patterns — including Malay, Manglish, English, Chinese, Cantonese, and code-switched conversations commonly used across the region. We support AI teams, startups, and research organizations with scalable human-in-the-loop workflows for: RLHF (Reinforcement Learning from Human Feedback) LLM response evaluation and ranking Conversational AI annotation Multilingual text labeling Sentiment and intent classification AI safety and cultural localization review Image and vision annotation Data collection and quality assurance Our workflow combines structured QA processes, multilingual reviewers, and AI-assisted operational tools to improve consistency, speed, and scalability across annotation pipelines. We specialize in areas where regional language and cultural understanding are critical, including conversational AI, localization, customer support systems, content moderation, and generative AI evaluation. Based in Southeast Asia, XTECQ is positioned to support the growing demand for high-quality multilingual datasets and culturally aligned AI evaluation workflows across emerging markets. We are currently focused on pilot collaborations and scalable AI data partnerships with startups, AI labs, and organizations building multilingual AI products.

Security

Security Overview

XTECQ adopts a security-conscious and confidentiality-focused approach to protect client data, project materials, and annotation workflows. Our operational workflows are designed with controlled data access, confidentiality practices, and secure collaboration standards suitable for AI data and annotation projects. Security measures include: Restricted access to project files and datasets based on project scope and personnel requirements Secure cloud-based collaboration environments with access permission controls Use of password-protected systems, encrypted communication platforms, and secure authentication practices Antivirus protection, firewall-enabled network environments, and regularly updated software systems NDA and confidentiality-based collaboration policies for internal contributors and project participants Controlled handling of sensitive datasets and client information Separation of project resources to minimize unauthorized access risks For workforce and operational security: Team members and collaborators are expected to follow confidentiality and responsible data handling guidelines Access to client data is limited only to authorized personnel involved in the project Annotation and QA workflows are reviewed to maintain consistency, accuracy, and controlled data exposure XTECQ is committed to continuously improving operational security practices as the company scales and supports larger AI data and multilingual annotation projects.

Labeling Experience

Multilingual Conversational AI Evaluation Workflow

Doccano

Text

RLHF

Classification

Designed and tested a multilingual conversational AI evaluation workflow focused on Southeast Asian language and cultural contexts. The project involved creating structured annotation and QA processes for conversational datasets containing English, Malay, Chinese, and code-switched dialogue samples. Tasks included: Response quality evaluation Intent classification RLHF-style response ranking Toxicity and safety review Data quality assurance The workflow was designed to support scalable human-in-the-loop AI evaluation processes for conversational AI and multilingual LLM applications. Quality measures included structured review guidelines, consistency checks, and manual QA verification across multilingual samples.

2026 - 2026