Hamsika Puthiamuthu

I’m an AI/LLM analyst aspirant who enjoys turning messy inputs into reliable signals. I focus on prompt annotation, LLM response evaluation, and dataset validation so AI outputs are accurate, consistent, and safer to use.

As a Data Annotator with Innodata (Freelancer), I annotate and validate large-scale datasets for AI/ML models. I follow detailed annotation guidelines, run quality checks, and maintain consistency across text/image/video data—especially under tight deadlines to keep model training moving fast.

Earlier, I interned as a Data Science Intern at iFenSys InfoTech Pvt. Ltd. I automated data processing pipelines with Python (Pandas, NumPy) to handle 2,000+ data points, cutting manual effort by 15%. I also built predictive workflows and used Matplotlib to create clear visualizations for stakeholder-ready reporting.

My projects reflect how I work in practice: I built an LLM Response Evaluation & Annotation Platform to score relevance, fluency, toxicity, and instruction-following using transformer-based NLP quality analysis. I also developed a Prompt Annotation & Quality Review System to classify prompts, validate responses, and produce structured summary reports—so evaluation insights remain traceable and actionable.