AI Research Engineer (Model Compression & Quantization)
Tether Operations Limited5 months ago
Remote, WorldwideSenior
Responsibilities
- Apply low-bit quantization to LLMs, VLMs, and other generative AI models while preserving accuracy and output quality.
- Use knowledge distillation to transfer capabilities from large teacher models to smaller multimodal student models.
- Implement pruning methods to remove redundant parameters and attention heads while maintaining task performance.
- Evaluate trade-offs among model size, latency, memory use, computational cost, and accuracy across compression methods.
- Research and apply mixed-precision quantization, adaptive pruning schedules, intermediate feature matching, and other advanced compression strategies.
- Build robust compression pipelines and establish performance and fidelity metrics for efficient inference.
- Identify and address production inference bottlenecks for low-memory, low-latency deployment on edge devices.
- Stay current with model compression research for multimodal and generative architectures.
- Document methodologies, experiments, and results to support reproducibility, collaboration, and stakeholder communication.
- Author technical papers and publish model-compression research at conferences such as NeurIPS, ICML, ICLR, CVPR, ACL, and AAAI.
Requirements
- Degree in Computer Science or a related field.
- Ideally, a PhD in NLP, Machine Learning, or a related field, with a strong AI R&D track record and publications at top-tier or A* conferences.
- Hands-on experience with PyTorch or an equivalent deep learning framework.
- Hands-on experience with both Quantization-Aware Training (QAT) and Post-Training Quantization (PTQ).
- Research and practical experience using knowledge distillation to compress large models into smaller, efficient models.
- Research and practical experience using model pruning to compress large models into smaller, efficient models.
- Strong understanding of neural network architectures and training processes, including transformers, LLMs, VLMs, backpropagation, optimization, and fine-tuning.
- Familiarity with C++ is preferred, particularly for low-level quantization kernels or inference optimizations.
Benefits
- Remote work from locations around the world as part of a global team.
Categories
AI ResearchML Engineering
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.