SLM Training Engineer
Core
Train, fine-tune, and deploy small language models (SLMs) and decision models to make enterprise AI agents reliable, affordable, and efficient at scale for industrial domains.
Role type
Senior IC machine-learning engineer (SLM and decision model training)
Builds
Specialized SLMs for tool selection and extraction; decision models for routing, triage, guardrails, and automated evaluation of agent outputs.
Domain
Industrial automation and manufacturing AI
Deliverable
production ML models
Required skills
SLM fine-tuning (LoRA, QLoRA, DPO), decision model training (calibrated probabilities), data pipelines (synthetic data, distillation), model evaluation (calibration, latency, cost), model optimization (quantization, pruning), MLOps (experiment tracking, monitoring)
Preferred skills
Experience with System One decision models (e.g., TypeSafe Jev), reinforcement learning for calibrated decisions, edge/industrial deployment, cloud platforms (Azure, AWS, GCP)
Technologies
PyTorch, Hugging Face Transformers, TRL, DeepSpeed, vLLM, ONNX Runtime, Triton Inference Server, MLflow, Azure Machine Learning
Responsibilities
Train and distill SLMs for specific industrial tasks; build decision models returning typed choices and calibrated probabilities; design evaluation suites comparing SLMs, decision models, and LLMs; optimize models for edge and on-premises deployment; manage the model lifecycle with MLOps practices; partner with business teams to identify bounded decision opportunities.
Seniority
Senior, hands-on IC
