Farhan Azim

Key Skills

Software

Data Annotation Tech

Toloka

Telus

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Video

Top Task Types

Action Recognition

Bounding Box

Computer Programming/Coding

Data Collection

Text Generation

Freelancer Overview

•Engineered high-quality datasets for training and fine-tuning LLMs and action agents using both supervised and unsupervised techniques. •Designed data labeling, cleaning, augmentation, and validation pipelines to ensure dataset consistency and reliability. •Created synthetic datasets for domain-specific and safety-critical scenarios, enhancing model robustness. •Collaborated with ML researchers to implement post-training workflows such as SFT, RLHF, and evaluator datasets. •Applied Python (Pandas, NumPy, PyTorch DataLoader) for data preprocessing, feature extraction, and batch pipeline optimization. •Contributed to annotation guidelines and documentation to streamline RLHF and model alignment processes.

Labeling Experience

Operations AI - AI Tutor

LabelboxVideoBounding BoxPolygon

Training General Agents, LLM Models, LAM General Agents like ACE. • Engineered high-quality datasets for training and fine-tuning LLMs and action agents using both supervised and unsupervised techniques. • Designed data labeling, cleaning, augmentation, and validation pipelines to ensure dataset consistency and reliability. • Created synthetic datasets for domain-specific and safety-critical scenarios, enhancing model robustness. • Collaborated with ML researchers to implement post-training workflows such as SFT, RLHF, and evaluator datasets. • Applied Python (Pandas, NumPy, PyTorch DataLoader) for data preprocessing, feature extraction, and batch pipeline optimization. • Contributed to annotation guidelines and documentation to streamline RLHF and model alignment processes.

Training General Agents, LLM Models, LAM General Agents like ACE. • Engineered high-quality datasets for training and fine-tuning LLMs and action agents using both supervised and unsupervised techniques. • Designed data labeling, cleaning, augmentation, and validation pipelines to ensure dataset consistency and reliability. • Created synthetic datasets for domain-specific and safety-critical scenarios, enhancing model robustness. • Collaborated with ML researchers to implement post-training workflows such as SFT, RLHF, and evaluator datasets. • Applied Python (Pandas, NumPy, PyTorch DataLoader) for data preprocessing, feature extraction, and batch pipeline optimization. • Contributed to annotation guidelines and documentation to streamline RLHF and model alignment processes.

2024 - 2025

Education

Work History