Machine Learning Engineer (LLM Infrastructure) - Toronto

at  Prodigy Labs

We are looking for a Machine Learning Engineer who has strong experience in building systems that accelerate the development and deployment of machine learning models, especially large language models (LLMs). You will partner closely with Machine Learning researchers and internal users to understand requirements and apply strong ML fundamentals to build high performance and reusable APIs and can also apply them in real production settings.

REQUIREMENTS

• 5-6 years of AI, Big Data and cloud expertise
• 3-4 years of Alternative data experience
• 2+ years of experience building machine learning training pipelines or inference services in a production setting
• Experience with LLM deployment, fine tuning, training, prompt engineering, etc
• Experience with LLM inference latency optimization techniques, e.g. kernel fusion, quantization, dynamic batching, etc.
• Experience with CUDA, model compilers, and other model-specific optimizations
• Experience building, deploying, and monitoring complex microservice architectures.
• Degree in Computer Science or Engineering
• Prior Experience with: -Docker, Kubernetes, Infrastrure as code (Terraform)and containerization, Agile Methodology, Distributed systems, Databricks ML, Cloud (Azure (preferred) or AWS)
• Expert level – Python, SQL
• Experience (or knowledge of) Mosaic ML, Ray Framework
• Experience with Lang Chain or LlamaIndex
• Experience with any vector database

• Architect/Enable distributed compute aligning workloads to Small/Mid/High end GPUs
• Leverage appropriate storage hardware and data formats to improve read/re-read efficiency
• Identify and remediate latency contributors especially IO bottlenecks, inefficient data shuffling, under/over utilized compute
• Scale models by employing distributed training using Data / Model Parallelism techniques
• Parallelize inference processing to improve prediction latency.
• Provide Subject Matter Expertise in Graph and Vector databases for various use cases including Knowledge Graphs, RAG etc.
• Implement LLM observability and monitoring solutions
• Manage infrastructure and large-scale system design and diagnose both model and system failures
• Mitigate reputation risk through AI driven Data Quality to ensure highest quality data and services are offered to clients