Senior ML & Inference Engineer — Toronto (Onsite)

Randstad; Toronto; On-site; CAD 90,000 - 130,000; Full time

Benefits offered by this job

Onsite in Toronto

7-month contract

Potential extension

Mentorship in ML/AI

Job summary

Randstad Canada in Toronto is seeking a Senior Software Developer for a 7-month onsite contract to lead AI/ML engineering, LLM pipelines, and model inference workloads.

You will run batch inferences in Databricks and cloud ML environments, work with DSPy, LangChain, and Hugging Face, and ensure reproducible data science workflows with clean notebooks and robust experimentation tracking.

Qualifications

• Degree in CS/DS/AI or related field.
• Strong Python and data science libraries experience.
• Experience with DSPy, LangChain, Transformers.
• Experience with Databricks/MLflow.

Responsibilities

• Execute and monitor batch inference jobs in Databricks and cloud ML environments.
• Troubleshoot inference issues, log results, and address model output anomalies.
• Configure, test, and run pipeline workflows using DSPy and related LLM frameworks.
• Load, test, and validate model weights, tokenizers, and configuration files.
• Support model evaluation runs using internal benchmarks and metrics.
• Track, record, and summarize performance across experiments.
• Prepare, preprocess, and clean text datasets for inference pipelines.
• Annotate and review model outputs for error analysis and benchmarking.
• Maintain clean, well-documented notebooks and Python scripts.
• Collaborate with engineering to ensure data handling and model versioning best practices.

Skills

Python

Pandas

Scikit-learn

PyTorch

LLM frameworks

Education

Degree in CS/DS/AI

Tools

Databricks

MLflow

DSPy

LangChain

Hugging Face

Job description

Randstad Canada in Toronto is seeking a Senior Software Developer for a 7-month onsite contract to lead AI/ML engineering, LLM pipelines, and model inference workloads.

You will run batch inferences in Databricks and cloud ML environments, work with DSPy, LangChain, and Hugging Face, and ensure reproducible data science workflows with clean notebooks and robust experimentation tracking.