Software Engineer

About the Job :

We are at the forefront of AI and machine learning, and we’re looking for motivated individuals to contribute to the next generation of intelligent models. The ideal candidate will have experience working with Turing.com and a strong background in data annotation, prompt engineering, and model fine-tuning. You will play a critical role in refining AI systems, providing essential human feedback, and enhancing overall model performance.

Responsibilities :

Develop quality software and web applications
Analyze and maintain existing software applications
Design highly scalable, testable code
Discover and fix programming bugs

Key Qualifications

Experience in RLHF:

Deep understanding of Reinforcement Learning (especially RLHF) for LLMs and how it applies to improving AI models.
Hands-on experience in fine-tuning LLMs through iterative human feedback.

Data Expertise:

Prior experience in annotating datasets for AI/ML models with a focus on quality control.
Experience with annotation tools and platforms like Labelbox, Prodigy, or Turing’s proprietary tools.

Technical Proficiency:

Familiarity with LLM frameworks like GPT-3/4, BERT, and advanced NLP models.
Strong command of Python, SQL, or related programming languages for handling data processing tasks.
Understanding of prompt engineering, and experience with platforms like Hugging Face or LangChain is a plus.

Preferred Qualifications:

Experience in model fine-tuning, prompt engineering, and human-in-the-loop systems.
Familiarity with cloud platforms (AWS, Azure) and MLOps best practices.
Previous work on reinforcement learning pipelines in large-scale AI projects.

What You’ll Do :

You will play a key role in annotating and curating data for the training and fine-tuning of large language models (LLMs), ensuring annotations are accurate, consistent, and project-aligned. You’ll implement Reinforcement Learning with Human Feedback (RLHF) techniques, providing structured human feedback to guide model outputs and continuously fine-tune models to improve performance.