Engineer transformer-based threat detection with optimized inference and dedicated GPU resources in a fully remote Canadian position. Scale models responsibly. This senior engineering role combines transformer expertise with effective machine learning for adversarial security environments. You will prepare datasets, fine-tune models, evaluate production risks, and deploy controlled releases. The focus includes CPU, memory, latency, throughput, and cost per file across large hosting fleets. Key Responsibilities:
Build reusable training pipelines and dataset tooling Use GPU resources for parallel experimentation Move models through shadow deployments and controlled releases Monitor detection rates and model performance trends Document architecture, limitations,
hardware, and production results Requirements:
Practical transformer training beyond hosted model APIs Experience with LoRA, QLoRA, distillation, and quantization Ability to optimize CPU, memory, throughput, latency, and cost Familiarity with ONNX, llama.cpp, vLLM, or integer quantization Ownership from experimentation through production deployment Bring transformer engineering, GPU-aware experimentation, and inference optimization to production malware detection and continuous model improvement.