Somewhere in this firm a lawyer just breathed easier because of what you built

Position: Permanent

Work environment: Hybrid (4 days in-office)

Location: Toronto

Salary: Up to $175,000 base

Position Overview

Hammehr invites applications for the role of Applied AI Scientist on behalf of our client for a full-time permanent opportunity.

Success in this role will look like firm-specific AI tools that legal teams consistently use and trust. The client is focused on applying AI in ways that support deep legal thinking and protect their proprietary knowledge. The work environment prioritizes precision, accepts ambiguity, and values clear outcomes over hype.

What You'll Do

• Build AI tools that help lawyers process large volumes of legal material without burnout
• Translate open-ended legal questions into practical models that serve daily work
• Meet with legal teams, clarify goals, and guide them through technical decisions
• Ship incremental improvements while maintaining alignment with long-term goals
• Troubleshoot bugs that arise from applying AI to legal and procedural text
• Treat accuracy, explain ability, and model risk as core to design, not afterthoughts
• Identify inefficient processes and replace them with cleaner, tool-based solutions
• Adjust and retrain models when outputs shift away from legal expectations

Basic Qualifications

• 3+ years of production-level Python experience with comfort owning model lifecycle and debugging
• Direct experience with LLMs or generative AI applied to live, domain-specific problems
• Familiarity with deploying AI systems in Azure or similar cloud environments
• Ability to explain technical decisions in practical terms to legal or business teams

Preferred Qualifications

• Prior work on projects that required compliance with legal or regulatory standards
• Experience handling complex or sensitive data, especially in legal, financial, or healthcare domains
• Built internal tools that continued to see use across multiple cycles or teams
• Conducted adversarial testing or explored model failure modes in sensitive applications
• Presented AI risks and limitations to decision-makers with varied technical backgrounds

The Challenges

• Translate between legal reasoning and machine learning without relying on shared vocabulary
• Work with inconsistent and fragmented data sources while aiming for dependable tools
• Design automation in an environment shaped by traditional metrics like billable hours
• Guide non-technical stakeholders through decisions about scope, design, and deployment

Your Impact

• Free legal professionals from repetitive review work and increase their cognitive focus
• Deliver tools that provide direct, actionable answers to pressing legal questions
• Embed institutional expertise into secure AI systems that reflect how the firm operates