Mitchele Jebet

Detail-oriented AI and data specialist with experience evaluating AI-generated content, annotating and labeling data, reviewing outputs for accuracy and consistency, and producing structured written documentation. Skilled at following complex guidelines precisely, identifying inconsistencies and edge cases, and maintaining quality standards independently across high-volume tasks. Strong written English, analytical thinking, and a demonstrated commitment to accuracy — with a technical background that allows evaluation of AI outputs in engineering, computing, and cybersecurity domains beyond surface-level language review.

AI Content Evaluation & Data Annotation at Independent Projects (2023-01 – Present)

• Built an AI Model Training & Expert Evaluation Framework — reviewed AI-generated English explanations for technical accuracy, clarity, completeness, and practical value using a consistent 1-5 scoring rubric across Electrical Engineering and Network Systems domains
• Identified errors, misconceptions, and poor phrasing in AI responses — provided corrected versions with documented reasoning and educational notes, producing comprehensive JSON training reports with all evaluations and summary statistics
• Built an AI Data Annotation & Quality Assessment Tool — systematic labelling with confidence scoring, multi-criteria quality checks (completeness, format, consistency), pairwise comparison, and CSV/JSON report generation; achieved 95.67% average quality score across test batches
• Annotated Swahili and mixed-language narratives for NLP classification — developed annotation protocols, documented edge cases, and maintained consistency standards across batches

Software Developer & Technical Reviewer at Self-Employed (2023-01 – Present)

• Reviewed AI-generated security analysis outputs for consistency, clarity, and technical accuracy — refined findings and recommendations to improve quality before presentation, using structured evaluation rubrics
• Tested software features against expected behaviour, documented defects with reproduction steps, and verified fixes — following structured testing procedures to ensure reliable and consistent application performance
• Built a 16-endpoint production REST API with 100% pytest test coverage and GitHub Actions CI/CD — all code reviewed, tested, and documented to professional standards
• Produced technical documentation including API guides, vulnerability disclosure reports, and forensic investigation write-ups — all in clear, precise English for technical and non-technical audiences
• Identified and responsibly disclosed two CVE-class vulnerabilities in production AI systems — documented with full reproduction steps, impact assessment, and remediation guidance

Professional Cybersecurity Analysis in Cybersecurity – Moringa School (2025-01 – 2025-12)