We’re looking for a talented AI Engineer with deep expertise in Large Language Models (LLMs) to join our team. This role will focus on designing, training, fine-tuning, and deploying state-of-the-art LLMs, as well as building scalable systems around them. If you’re passionate about pushing the limits of generative AI and building real-world applications, we’d love to hear from you. Responsibilities
Research, design, and implement solutions leveraging LLMs and LLM APIs.
Evaluate, train, fine-tune, and optimize open-source LLM models.
Develop robust prompt engineering and prompt optimization strategies.
Implement and compare different LLM training methods (supervised fine-tuning, RLHF, LoRA, quantization, etc.).
Build scalable backend services integrating LLMs into products.
Collaborate with product and engineering teams to design AI-powered features.
Stay up to date with the rapidly evolving AI/LLM ecosystem.
Requirements
Strong experience with LLMs: understanding of model architectures, APIs, training/fine-tuning, and optimization.
Hands-on experience with open-source models and modern AI frameworks (PyTorch, Hugging Face, etc.).
Expertise in prompt design and evaluation.
Solid programming skills in TypeScript (experience with Python is a plus).
Knowledge of cloud platforms, preferably Google Cloud Platform (GCP) + VertexAI, for deploying and scaling AI workloads.
Strong understanding of distributed systems and performance optimization for ML.
Excellent problem-solving skills, curiosity, and ability to work in a fast-paced environment.
Nice-to-Haves
Experience with vector databases, embeddings, and retrieval-augmented generation (RAG).
Familiarity with MLOps pipelines and deployment best practices.
ActionModel is building the next generation of AI — one that can use the internet like a human. Our intelligent agents interact directly with websites to complete real tasks: clicking, typing, navigating, and executing workflows across any platform.
We’re pioneering a new layer of automation that works through the user interface, not just APIs — unlocking the full potential of AI-powered productivity. By learning from real human behavior, the ActionModel gets smarter with every interaction.