Open to opportunities

ashish kumar

@ashishkumar63

Message

I evaluate multimodal AI outputs and build accurate training datasets.

India

Message

What I'm looking for

I'm looking to contribute to multimodal AI and GenAI projects where I can improve training-data quality, evaluate model outputs, and apply rigorous annotation and validation standards.

I have experience working on AI training and data annotation projects involving image, video, text, and voice data. My work included video annotation, frame-by-frame review, video segmentation, timestamping, object and action annotation, and annotation verification. I also worked on bounding boxes, semantic segmentation, and checking annotations against detailed project guidelines.

I have also worked on GenAI evaluation, where I reviewed AI-generated responses for accuracy, relevance, naturalness, instruction following, and consistency using given rubrics. I regularly performed quality checks, identified errors, and corrected or flagged them before submission.

I have used tools such as CVAT, Labelbox, SuperAnnotate, and other annotation platforms. I have not worked directly on robotics training, but I have relevant experience with video and frame-level annotation for AI training.

Experience

Work history, roles, and key accomplishments

Current

Data Annotator

Current

Innodata

Jan 2026 - Present (8 months)

Contributed to 6+ multimodal AI and GenAI projects across image, video, text, and voice modalities. Performed annotation, evaluation, and quality assurance tasks maintaining high accuracy and SLA adherence.

Bounding Box Annotation Object Tracking Semantic Segmentation Dense Labeling Annotation Validation Generative AI Evaluation RLHF Style Evaluation AI Output Evaluation Dataset Validation Data Quality Assurance Video Segmentation Referring Expressions Superannotate CVAT Labelbox Excel SQL Python Slack

Data Annotator

Micro1

May 2025 - Jul 2025 (2 months)

Processed and structured image, text, and video datasets for ML and AI training. Performed bounding box, polygon, and semantic segmentation annotation using Labelbox, CVAT, and SuperAnnotate.

Bounding Box Annotation Polygon Annotation Semantic Segmentation Dataset Validation Data Quality Assurance Frame By Frame Analysis Labelbox CVAT Superannotate

Associate Trainee Recruiter

Cogent Infotech

Jul 2022 - Oct 2022 (3 months)

Screened and evaluated candidate profiles for US-based IT positions. Conducted inbound and outbound calls, managed candidate records, and updated ATS databases.

Candidate Screening Recruitment ATS Data Integrity Communication Skills