Moiz
Level 1
Habilidades
Revisa mis servicios


¿Quieres trabajar por horas?
Dile a Moiz qué necesitas.
USD 15
/
horaPorfolio
Experiencia laboral
LLM Engineer
Microsoft • Tiempo completo
Sep 2024 - Dec 2025 • 1 yr 3 mos
Designed and developed a Human-in-the-Loop (HITL) evaluation platform to validate and improve Agentic AI function calling capabilities. This tool serves as a critical component in the RLHF (Reinforcement Learning from Human Feedback) pipeline. - Core Workflow: Built a quiz-style interface where human graders analyze a User Query versus the Assistant’s Tool Usage. - Granular Evaluation: Implemented a rigorous 9-10 question framework allowing graders to assess specific metrics, including tool selection logic, parameter accuracy (groundedness), and overall response completeness. - Data Pipeline: The platform captures granular feedback to generate high-quality labeled datasets used to fine-tune models, ensuring the AI strictly adheres to function definitions and user intent.
28 Reseñas
| (27) | ||
| (2) | ||
| (0) | ||
| (0) | ||
| (0) |
Desglose de calificaciones
- Nivel de comunicación del Freelancer
- Calidad de la entrega
- Valor de la entrega
Ordenar por
moez_j

Canadá
He demonstrated outstanding professionalism in his approach to work, and it was clear that he was deeply invested in producing a final result that met our goals.

kofworkshard

Japón
Working with Moiz was an absolute pleasure. His professionalism and detailed documentation made the project run smoothly, and his deep understanding of software development truly stood out. Highly recommend! 👍

kofworkshard

Japón
This guy provided us with high-quality and efficient code throughout the project. He accurately understood our requirements and always suggested the best solutions, allowing the project to progress smoothly. We truly appreciate his excellent technical skills and problem-solving abilities."

wordpress7007

Bangladesh
krishnavari9

Estados Unidos