Senior Machine Learning Engineer, Classification and Detection

Smartverify•,

17h•CA$120,000 - CA$160,000•Hybrid

About The Position

This role focuses on training models for classification and detection systems from scratch, including fine-tuned transformers, entity detection over structured and unstructured content, and behavioral scoring. It is distinct from roles focused on retrieval-augmented generation, prompt engineering, agent orchestration, or integrating foundation model APIs. The ideal candidate will have experience owning a classifier in production, covering the entire lifecycle from data preparation through training, evaluation, and retraining. SmartVerify is developing the data egress control plane for enterprise AI, positioned between AI agents and enterprise data to inspect queries, enforce policies in real-time, and create an immutable audit trail. This is a greenfield project based on an existing specification, with the opportunity to refine the spec. The role involves being the sole Machine Learning Engineer, reporting directly to the founder, with support from a co-op for evaluation harness and pipeline QA.

Requirements

Production ML ownership: experience training, deploying, monitoring, and retraining models in a live system.
Transformer fine-tuning for text classification (BERT family, DistilBERT, or equivalent).
Sequence labelling or named entity recognition for structured entity detection.
Proficiency in Python, PyTorch, and Hugging Face Transformers.
Demonstrated evaluation rigor: ability to explain threshold choices and trade-offs.
Comfort working from a written design specification.
Authorized to work in Canada or the United States.
Located in British Columbia or the Seattle area.

Nice To Haves

Bootstrapping labels through weak supervision, LLM-assisted labelling, or active learning.
Anomaly or behavioral detection with sparse or absent labels.
Experience with SageMaker training jobs and inference endpoints.
Experience in regulated domains such as HIPAA, PCI DSS, GDPR, or SOC 2.
Background in fraud, abuse, or security detection.
Knowledge of distillation or quantization for inference latency budgets.
Experience with Kinesis, Kafka, or other streaming pipelines.
SQL and query structure parsing.
Mentoring a junior engineer or co-op.

Responsibilities

Owning the behavioral classification model: multi-class classification of AI agent query intent against an internal taxonomy, producing labels, confidence, and supporting evidence.
Implementing PII and PHI detection over query content and returned data, including span-level identification for audit evidence.
Developing behavioral risk scoring by combining deterministic request signals with model-derived signals.
Managing model training, serving, and versioning on SageMaker within the asynchronous inspection path.
Establishing the evaluation harness, drift detection, and retraining loop, with a co-op supporting the harness work.
Defining the classification output contract for downstream consumers (audit, enrichment, dashboard).