Job Description
Are you ready to shape the future of Generative AI? At Synthetix AI Labs, we are building the next generation of cognitive systems that redefine human-machine collaboration. As a Senior Machine Learning Engineer, you will join an elite, fast-paced R&D team designing, training, and deploying large-scale neural networks.
We provide a state-of-the-art compute infrastructure (thousands of H100 GPUs), a highly collaborative environment, and the opportunity to see your models impact millions of users globally. If you thrive on solving unstructured, frontier-level AI challenges, this is your next career defining move.
Responsibilities
- Lead the architecture, training, and fine-tuning of multi-billion parameter Large Language Models (LLMs) and diffusion models.
- Optimize distributed training pipelines across massive GPU clusters using PyTorch, DeepSpeed, and Megatron-LM.
- Implement advanced RLHF (Reinforcement Learning from Human Feedback) and DPO pipelines to align model behaviors.
- Collaborate with product and infrastructure teams to deploy low-latency, high-throughput inference APIs.
- Conduct pioneering research into novel attention mechanisms, state-space models, and vector database retrieval-augmented generation (RAG).
- Mentor junior machine learning engineers and contribute to our collaborative, high-performance engineering culture.
Qualifications
- Master's or Ph.D. in Computer Science, Mathematics, or a related quantitative field with a focus on Deep Learning.
- 4+ years of professional experience training and deploying deep learning models in production environments.
- Expert-level proficiency in Python and deep learning frameworks, specifically PyTorch.
- Proven track record of scaling distributed training workloads across multi-node GPU environments.
- Deep mathematical understanding of transformer architectures, optimization algorithms, and NLP/Computer Vision fundamentals.
- Experience with modern MLOps tools (e.g., Weights & Biases, Kubernetes, Triton Inference Server).
- First-author publications in top-tier AI conferences (NeurIPS, ICML, ICLR, CVPR) is a strong plus.