Machine Learning Engineer

Let’s start with some numbers:

• H100 SXM: 504 (SuperPod)
• H100 PCI: 128
• A100 SMX: 120
• CPU: 600K cores
• 600k CPU Cores
• 350pb SSD
• 33pb HDD

That is a lot of computing power, so why do we need it? 

We’re pushing the boundaries of AI by building cutting-edge machine learning models and tools from the ground up. To support this, we’ve invested heavily in high-performance computing, including the Ahrefs Superpod, which boasts 60 nodes from 63 and scored 19.7 PFlops in Linpack, ranking it among the top 30 supercomputers in the world. Simply put, we have the infrastructure to turn ambitious ideas into reality.

For over a decade, Ahrefs has been crawling the web, processing petabytes of data, and delivering powerful insights through an intuitive interface. Now, we’re leveraging this expertise to develop innovative AI-driven solutions that help users navigate, analyze, and extract value from vast amounts of information like never before.

Let’s put that compute power to use! 

What You’ll Do

In this role you’ll have the opportunity to work on cutting-edge research and turn it into real-world impact. We're looking for a dynamic Machine Learning Engineer to join us in creating cutting-edge AI products that push the boundaries of what’s possible. Bring your creativity, develop your own ideas, and see them through, from fuzzy concepts to production, all in a fast-moving, red-tape-free environment.

• Model Development
  • Work on LLMs and other AI models, spanning training, fine-tuning, inference optimization, and new architectures.
• Algorithm Research & Development
  • Design and implement algorithms for search, agents, knowledge retrieval, and generative AI applications.
• Stay on the Cutting Edge
  • Stay ahead of NLP and RL advancements, evaluate emerging models/frameworks (e.g., SSMs, text diffusion), and assess their impact on our products.
• Scaling & Optimization
  • Optimize model inference, reduce latency, and improve efficiency using techniques like distillation, quantization, and tensor parallelism.
• Collaborate with Engineering & Product Teams
  • Work closely with engineering and product teams to ship AI-driven features at scale.

What We’re Looking For

• Experience with modern deep learning frameworks (e.g., PyTorch, JAX).
• Familiarity with transformer-based architectures and techniques (e.g. flash attention, speculative decoding, KV caching).
• Strong software engineering skills (Python, CUDA, or C++ a plus).
• Interest in one or more areas: large-scale training, agentic workflows, inference optimization, retrieval-augmented generation (RAG), reinforcement learning (RLHF/RL), or multimodal AI.
• Experience with distributed training and model deployment libraries (e.g. Megatron, Ray, Triton (both or either!), vLLM) is a bonus.

You’ll have opportunities to work on projects at the cutting edge of AI. You’ll get the chance to shape the future of our product and make a real impact.

If you're a Machine Learning Engineer with the drive to turn research into reality, we'd love to hear from you! Join us and be part of a team that's pushing the boundaries of technology and innovation.