Shahnawaz Ahmed

Key Skills

Software

AWS SageMaker

CVAT

Doccano

Google Cloud Vertex AI

Labelbox

Mindrift

Scale AI

Snorkel AI

SuperAnnotate

Supervisely

Label Studio

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Text

Top Task Types

Evaluation Rating

Fine Tuning

Prompt Response Writing SFT

Question Answering

RLHF

Freelancer Overview

AI Trainer and LLM Evaluation Specialist with hands-on experience in Reinforcement Learning from Human Feedback (RLHF), prompt engineering, and multi-modal annotation. Worked on projects for Amazon Phoenix, Google Gemini, and Fika GenAI via Invisible Technologies, Outlier AI, and Mindrift. Skilled in side-by-side evaluation, hallucination detection, and reward-model training data curation that improve factuality, safety, and alignment in LLMs.

Labeling Experience

Invisible Technologies — Amazon Phoenix (RLHF 3H Framework)

Label StudioTextQuestion AnsweringRLHF

Evaluated large volumes of model-generated responses under the Honest–Helpful–Harmless (3H) framework for Amazon’s Phoenix RLHF pipeline. Applied hallucination detection, bias and safety screening, and factuality verification across text-based dialogue and Q&A tasks. Authored structured preference rationales that clarified trade-offs between helpfulness and safety. Contributed to a measurable ~40% improvement in factual accuracy and reduced unsafe outputs across downstream LLM deployments.

Evaluated large volumes of model-generated responses under the Honest–Helpful–Harmless (3H) framework for Amazon’s Phoenix RLHF pipeline. Applied hallucination detection, bias and safety screening, and factuality verification across text-based dialogue and Q&A tasks. Authored structured preference rationales that clarified trade-offs between helpfulness and safety. Contributed to a measurable ~40% improvement in factual accuracy and reduced unsafe outputs across downstream LLM deployments.

2023 - 2024

Education

Work History