Open to opportunities

Mohammad Khurram

@mohammadkhurram

Message

I evaluate AI outputs, datasets, and code for quality, safety, and instruction compliance.

India

Message

What I'm looking for

I'm looking for AI evaluation, AI safety, data annotation, or quality assurance work where I can assess model outputs, validate datasets, apply detailed guidelines, and help improve Generative AI quality.

I'm currently evaluating AI-generated responses with Mindrift and Toloka, reviewing factual accuracy, coding correctness, reasoning, formatting, and instruction following across Generative AI projects.

I've performed RLHF evaluation, preference ranking, prompt assessment, and response comparison, while identifying JSON issues, annotation inconsistencies, linguistic errors, and guideline violations. I work across multilingual text, image, video, and audio datasets and document clear quality findings to improve model outputs.

As a Hindi AI Safety Evaluator with Mercor, I assess Hindi responses for safety, factual correctness, grammar, cultural appropriateness, and policy compliance. I identify unsafe, biased, inaccurate, or low-quality outputs using structured evaluation rubrics.

I bring a Computer Science Engineering background, native Hindi fluency, professional English communication, and practical knowledge of JavaScript, basic Python, JSON, REST APIs, Git, and data structures.

Experience

Work history, roles, and key accomplishments

Current

AI Quality Evaluation Specialist

Current

Mindrift, Toloka

Jan 2024 - Present (2 years 8 months)

Evaluated AI-generated responses for factual accuracy, coding correctness, instruction following, reasoning quality, formatting, and overall response quality. Reviewed AI outputs across multilingual text, image, video, and audio datasets while validating formatting consistency and structured outputs.