Amulya Tiwari

Key Skills

Software

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Fine Tuning

Evaluation Rating

Function Calling

Prompt Response Writing SFT

Freelancer Overview

I am a data analyst and prompt engineer with over 2.5 years of experience working across fintech, media, education, and AI domains. My expertise lies in data labeling, annotation, and the development of high-quality training datasets for machine learning and AI applications. I have hands-on experience annotating and labeling more than 5,000 images for computer vision projects, optimizing workflows using LabelImg, and evaluating AI-generated responses for accuracy and safety to improve model performance. My technical skills include SQL, Python, Power BI, Excel, and key data science libraries like Pandas, NumPy, and Matplotlib. I have contributed to projects involving medical image segmentation, e-commerce customer analytics, and the creation of synthetic datasets, consistently ensuring data quality and actionable insights for model development and deployment. I am passionate about leveraging data to drive reliable AI solutions and am committed to maintaining the highest standards in data annotation and training data processes.

Labeling Experience

Prompt engineer/AI labeller

LabelboxTextFine TuningEvaluation Rating

Designed YAML-based benchmarks to evaluate LLM reasoning under implicit constraints, conflicting user intents, and real-world workflow dependencies. Built simulated environments with explicit world state, action schemas, and execution rules to objectively test decision sequencing and outcome correctness. Developed and audited evaluation rubrics to eliminate hidden failure modes and reliably differentiate model performance (pass/fail) without unfair ambiguity.

Designed YAML-based benchmarks to evaluate LLM reasoning under implicit constraints, conflicting user intents, and real-world workflow dependencies. Built simulated environments with explicit world state, action schemas, and execution rules to objectively test decision sequencing and outcome correctness. Developed and audited evaluation rubrics to eliminate hidden failure modes and reliably differentiate model performance (pass/fail) without unfair ambiguity.

2025

Education

Work History