Patrick Selebunga

Contributed French-English speech and conversation recordings for AI training and evaluation, following linguistic and recording guidelines.

Analyzed Lingala (ln_CD) text and dialect variations for language identification, maintaining quality and consistency through guideline-based annotation.

Created and refined prompts; evaluated and ranked AI outputs for accuracy, consistency and user relevance, identifying errors and potential biases. Applied structured quality reviews and data cleaning, documenting training procedures and quality benchmarks.

Executed predefined test cases and exploratory tests; reported bugs with reproduction steps, severity and UX feedback.

Tested registration, payment flows and usability across desktop and mobile; documented defects and recommendations against product requirements.

Participated in AI community translation tasks, contributing to language data and AI training efforts.

Transcribed and reviewed Swahili audio for AI/ML language-data workflows, applying detailed transcription guidelines and quality standards. Checked linguistic accuracy and consistency, resolving ambiguities and transcription issues.

Captured real-world video for computer vision datasets, following technical specifications, privacy requirements and quality standards.

Created scripted human recordings, AI avatars and generated-video samples for media-integrity research, following privacy and consent guidelines.

Created prompts to assign tasks to AI agents in Fleet.ai; ranked LLM responses and validated outputs for clarity, relevance, user intent and guideline compliance. Reviewed French-language datasets and provided human feedback for RLHF.

Transcribed and annotated Lingala audio, applying project guidelines and self-review for linguistic accuracy.

Ensured data quality for AI training datasets, applying quality standards and guidelines.

Translated and localized French, English and Lingala content for accuracy, naturalness, consistency and cultural appropriateness.

Completed introductory coursework on agent engineering and approaches to more robust, production-ready AI agents.

Bachelor's Degree in Applied Computer Science for Business Management from ISIG GOMA, completed from 2023 to 2024.

Bachelor's Degree in Computer Science and Management from ISIG GOMA, completed from 2019 to 2022.