Data Scientist

Sonatype•Toronto, ON

About The Position

Sonatype is a leader in the software supply chain management industry, known for inventing componentized software development and pioneering the software supply chain category. They operate the world's largest repository of Java open-source components, Maven Central. Their platform helps organizations rapidly create, deploy, and maintain innovative software at scale, ensuring security and business alignment. Trusted by over 2,000 organizations, including 70% of the Fortune 100, and over 15 million developers, Sonatype's tools and guidance are essential for delivering exceptional, secure software. The company is committed to constant innovation, leveraging AI/ML to provide confidence in software quality, automation, and security. They have a history of innovation, from Nexus Repository to solutions that halt malicious open-source malware.

Requirements

5+ years of hands-on experience in applied data science, machine learning, AI engineering, or AI research.
Strong Python skills and practical experience with data and AI libraries/platforms such as Databricks, and LLM APIs, scikit-learn
Experience building and shipping ML or GenAI applications—from early prototype through usable internal or customer-facing workflows.
Deep familiarity with modern LLM ecosystems, including OpenAI, Anthropic/Claude, Hugging Face, and open-weight models.
Ability to select models and design effective LLM applications using prompting, context management, structured outputs, retrieval, and tool use.
Experience building agentic or multi-step AI workflows with LangGraph, LangChain, Semantic Kernel, or similar orchestration frameworks.
Strong evaluation mindset: defining useful quality metrics, building representative evaluation datasets, assessing reliability, and making data-driven tradeoffs.
Comfortable working with large, messy, structured, and unstructured data to produce features, insights, and clear visualizations.
Proficiency with Git, testing, code review, and collaborative software-development practices.
Practical, balanced judgment: comfortable exploring emerging AI capabilities while building maintainable, secure, dependable systems.
Proactive and accountable, with strong written and verbal communication skills across technical and non-technical partners.

Nice To Haves

Strong MLOps experience, including MLflow or comparable tooling, experiment tracking, reproducible pipelines, model/application versioning, CI/CD, serving, and production monitoring.
Experience operating ML or GenAI systems at scale, including observability, tracing, incident response, and data or model-drift detection.
Experience with Databricks ML, AWS SageMaker, Azure ML, or similar managed ML platforms.
Familiarity with MCP, agent-tool integrations, LLM guardrails, and production safety practices.
Experience with AI-assisted development tools such as Copilot, Claude Code, or Codex.
Exposure to cybersecurity, fraud detection, anomaly detection, code analysis, or software supply-chain security.
Experience with PySpark and production data pipelines.
Experience working within a software product company or SaaS.

Responsibilities

Lead applied AI projects from concept to impact – prototype, validate, and help teams deploy practical ML and GenAI solutions.
Act as an internal consultant across product, engineering, security, and research teams: scope problems, evaluate approaches, and advise on ML/AI best practices and productive use of generative technologies.
Lead the research, development, and deployment of models for use cases such as malicious behavior detection, anomaly detection, and fraud analysis – using techniques ranging from classical ML to LLMs, embeddings, retrieval-augmented generation, and agentic workflows.
Design robust experiments and establish evaluation pipelines for model reliability, accuracy, and business impact (cross-validation, drift monitoring, ground-truth evaluation).
Bridge research and production: translate research insights into scalable APIs, tools, or workflows that enable other teams to adopt AI effectively.
Explore new techniques (LLMs, embeddings models, RAG, agentic workflows) to enhance developer and security experiences.
Communicate technical concepts, tradeoffs, and recommendations clearly to both technical and non-technical stakeholders through presentations, documentation, and collaboration; mentor peers and help elevate the organization's AI literacy and capabilities.
Partner with our data governance team to ensure compliance with data-privacy regulations and ethical considerations when working with customer data.