AI Research Engineer (Model Compression & Quantization)
Tether Operations Limited5 months ago
Remote, WorldwideSenior
Responsibilities
- Apply low-bit and mixed-precision quantization, including Quantization-Aware Training and Post-Training Quantization, to LLMs, VLMs, and other multimodal generative models.
- Use knowledge distillation to transfer capabilities from larger teacher models to smaller student models across text, image, and audio inputs.
- Implement pruning methods to remove redundant parameters and attention heads while preserving task performance.
- Evaluate trade-offs among model size, memory use, latency, throughput, and accuracy, and propose improvements based on empirical results.
- Develop advanced compression strategies such as adaptive pruning schedules and distillation with intermediate feature matching.
- Build robust compression pipelines and address production inference bottlenecks for efficient edge deployment.
- Establish performance and fidelity metrics and document methodologies, experiments, and results for reproducibility and collaboration.
- Track developments in multimodal and generative model compression and publish technical findings in top-tier conferences.
Requirements
- Bachelor’s degree in Computer Science or a related field is required; a PhD in NLP, Machine Learning, or a related field is preferred.
- Solid track record in AI research and development, ideally including publications in A* conferences.
- Hands-on experience with model quantization, including both Quantization-Aware Training and Post-Training Quantization.
- Research and practical experience with knowledge distillation for compressing large models.
- Research and practical experience with model pruning for compressing large models.
- Strong understanding of neural network architectures and training, including transformers, backpropagation, optimization, and fine-tuning.
- Experience with PyTorch or an equivalent deep learning framework.
- Familiarity with C++ is a plus, particularly for low-level quantization kernels or inference optimizations.
Benefits
- Remote work with a globally distributed team.
About Tether Operations Limited
Tether Operations Limited develops and issues reserve-backed digital tokens used by exchanges, wallets, payment processors, and merchants, most notably the USDT stablecoin available across multiple blockchains. Founded in 2014 and privately held, it earns revenue from token issuance, reserve management, and related services, and has expanded into tokenization, peer-to-peer communications (Keet), and Bitcoin mining. USDT is widely used in crypto trading and cross-border transfers, including in emerging markets.