Kunal Pidiha

Detail-oriented AI Evaluation Engineer and Data Annotator with 2+ years of experience labelling, categorizing, and quality-assessing multimodal datasets, including image, video, audio, and text for AI/ML training pipelines. Proven ability to follow structured annotation guidelines with precision and consistency, collaborate with QA and project management teams, and deliver high-accuracy outputs under a deadline. Multilingual (Hindi native, English fluent) with a strong STEM background in Computer Science and hands-on experience improving AI model performance through data quality.

LLM Evaluation Engineer | AI Systems & Model Quality at Invisible (2024-05 – Present)

• Annotated, labelled, and categorised multimodal AI datasets, including image, video, audio, and text, against defined quality guidelines to ensure accuracy, consistency, and reliability.
• Designed and executed structured AI annotation and quality assurance workflows, maintaining high labelling precision across diverse data types and project requirements.
• Evaluated AI model outputs across modalities, providing detailed, actionable feedback to improve model performance and annotation consistency.
• Collaborated with engineering, QA, and project management teams to meet annotation deadlines, refine labeling guidelines, and uphold quality standards.
• Reviewed and validated autonomous AI agent outputs; defined gold-standard benchmarks and maintained annotation consistency across complex, ambiguous datasets.
• Applied domain-specific annotation techniques to image and video datasets, identifying edge cases, labelling errors, and inconsistencies to strengthen model training pipelines.
• Leveraged C++, JavaScript, React, and Kotlin to implement and debug AI-related annotation task pipelines across multi-language development environments

AI Software Developer at Invisible (2023-12 – 2024-05)

• Improved the accuracy, completeness, and reliability of AI training datasets through systematic annotation, structured labeling, and quality evaluation across image, audio, and text modalities.
• Applied advanced annotation and quality assessment strategies to benchmark and enhance model performance against defined standards, delivering consistent, reproducible results.
• Contributed to the design and refinement of annotation guidelines, enabling continuous quality feedback loops and measurable improvements in labeled data consistency.
• Handled domain-specific labeling projects requiring independent judgment on ambiguous or complex data samples with high attention to detail.
• Delivered scalable AI data solutions using Java, C++, JavaScript, and Python, supporting multilingual and multi-domain annotation projects.
• Applied data structures and algorithms expertise to resolve complex edge cases in AI data pipelines and annotation workflows.

Software Developer Intern at Labridge (2022-12 – 2023-02)

• Redesigned UI/UX interfaces across multiple applications, driving a 20% increase in user engagement through improved usability and responsive design.
• Streamlined API integration processes, eliminating data-flow bottlenecks and reducing API-related issues across production environments.
• Built responsive front-end applications using ReactJS and Material UI, delivering polished and accessible user experiences.