AI Engineer/ML Engineer - Senior Developers - AI Training - San Francisco, US

prolificacademicltd; United States; Remote; USD 55,000 - 110,000; Full time

Benefits offered by this job

Pay up to $80/hour

Fully remote engagement

Short skills assessment

Job summary

prolificacademicltd is seeking senior AI/ML engineers to shape the next generation of large language models through paid research contributions. This remote, task-based role lets you work on your schedule, bringing deep expertise to evaluation and alignment work.

It targets practitioners based in or near San Francisco for focused technical review. Responsibilities include evaluating model explanations, auditing code and notebooks, and providing RLHF feedback to support alignment.

Qualifications

• Advanced degree in ML with research or production experience.
• Experience building or fine-tuning ML models in production.
• Strong knowledge of Transformers, CNNs, RNNs, and optimization.
• Experience with RLHF or retrieval-augmented generation (RAG).
• Detail-oriented with ability to audit complex model logic.

Responsibilities

• Evaluate AI-generated explanations for technical correctness.
• Audit ML code and notebooks, including training loops and evaluations.
• Provide structured human feedback for RLHF and model alignment.
• Assess how models traverse chain-of-thought prompts for reasoning flaws.
• Run comparative benchmarks using specified metrics.

Skills

ML model deployment

RLHF

Prompt engineering

Mathematical reasoning

Attention to detail

Education

Master's or PhD in CS/AI

Tools

PyTorch

TensorFlow/Keras

AWS SageMaker

Weights & Biases

LangChain

Pinecone

Job description

Role overview

This participant-based role is open to senior AI and machine learning engineers interested in shaping the next generation of large language models through paid research contributions. Rather than a traditional hire, contributors complete flexible, remote tasks on their own schedule while bringing deep technical expertise to model evaluation and alignment work. It is designed for experienced practitioners based in or near San Francisco who can deliver focused, high-quality technical review.

Responsibilities

• Evaluate AI-generated explanations of model architectures, loss functions, and backpropagation for technical correctness.
• Audit machine learning code and notebooks, including training loops, preprocessing scripts, and model evaluations.
• Provide structured human feedback that powers reinforcement learning from human feedback (RLHF) and supports model alignment.
• Critically assess how AI models traverse complex chain-of-thought prompts and identify where reasoning fails.
• Run comparative benchmarks between model outputs using specified technical taxonomies and performance metrics.

Requirements

• Bachelor's, Master's, or doctoral degree in Computer Science, Artificial Intelligence, Robotics, or a closely related quantitative discipline with a machine learning focus.
• Proven experience building, deploying, or fine-tuning ML models in production environments.
• Strong grasp of neural network architectures such as Transformers, CNNs, and RNNs, paired with knowledge of optimization techniques.
• Practical experience with prompt engineering, RLHF, or retrieval-augmented generation (RAG) workflows.
• Keen attention to detail for detecting hallucinations, biased outputs, or logical flaws in AI-generated technical material.
• Ability to audit intricate model logic, assess mathematical proofs behind ML algorithms, and recognize training data contamination.

Nice to have

• Deep proficiency with PyTorch or TensorFlow/Keras.
• Advanced Python skills covering NumPy, Pandas, Scikit-learn, and Hugging Face Transformers.
• Experience with AWS SageMaker, Google Cloud Vertex AI, or tools like Weights & Biases and LangChain.
• Familiarity with vector databases such as Pinecone, Milvus, or Weaviate in RAG contexts.

Benefits and work setup

• Pay rates reaching up to $80 per hour for qualified participants.
• Fully remote engagement with flexible hours.
• Short 10 to 15 minute skills assessment as the entry point.
• Individual tasks usually fit within one uninterrupted hour, with many sessions shorter than that.