Sana

Client: Nvidia.

Associate Process Executive at RANDSTAD Pvt Ltd. (2024-05 – Present)

Client: Nvidia.

• Evaluated and benchmarked AI agent responses for accuracy, task completion, reasoning quality, and factual correctness through detailed fact-checking.
• Compared outputs from multiple AI systems (text, image, video) to assess performance, consistency, and alignment with prompts.
• Designed and refined prompts to improve model output quality, relevance, and overall performance.
• Performed image and video editing tasks including object addition, removal, replacement, and reconstruction using masking techniques.
• Assessed AI-generated audio for pronunciation, intonation, pacing, and clarity; also recorded and transcribed audio data as per guidelines.
• Trained AI models in object detection using structured hints, referring expressions, and evaluation of reasoning processes.
• Conducted quality analysis of reconstructed media (images/videos) by comparing with original data based on visual fidelity, aspect ratio, and accuracy.
• Skilled in image, video, and LiDAR data annotation, including 2D/3D labelling for object detection and autonomous driving applications.
• Strong expertise in semantic and instance segmentation, ensuring precise pixel-level annotations.
• Conducted quality assurance checks to maintain high accuracy, consistency, and adherence to guidelines.
• Experienced in large-scale, multi-class annotation projects, following best practices for AI/ML dataset preparation

M.TECH in Computer Science Engineering – Sharnbasva University Kalaburagi (2024)

B.TECH in Computer Science Engineering – Sharnbasva University Kalaburagi (2022)

SSLC (2016)