Legal AI Engineer

Legal AI Engineer

Telon is an AI-native services company. We embed lawyers and engineers directly into client teams to design, build and run AI-powered workflows, moving legal and operational work from manual process to intelligent automation. Our clients range from fast-growing businesses to global enterprises, and the work is delivered from hubs in the UK, US, South Africa and Argentina.

The role

Your core job is reading definitive agreements and judging what an AI model has made of them. On transfer and financing work, you go straight to the provisions that carry the risk and decide whether the model has read each one correctly: right, partially right or wrong, and why.

You are a qualified lawyer who knows what matters in a deal document and can hold a model to that standard across hundreds of agreements. You own the interpretation. You escalate the genuinely novel tax or regulatory questions and nothing else.

The first engagement is focussed on project finance and renewable energy, validating AI extraction across a large set of term sheets and definitive agreements. The role grows from there into configuring and tuning the review itself, and into Telon's wider legal engineering work on legal AI and contract platforms.

What you will do

• Read definitive agreements for the provisions that carry the risk in a transfer or financing: conditions precedent, indemnities and caps, survival, audit and information rights, change of control and termination triggers.
• Decide whether the model's answer on each provision is right, partially right or wrong, and record why, so that every call is auditable.
• Hold the line between interpretation and genuine novelty. You resolve the everyday judgment calls yourself and escalate only the new tax or regulatory questions.
• Run exception review on the clauses the model flags as low confidence.
• Configure and validate AI-assisted review-table extraction across large volumes of commercial agreements, and build and tune the clause-level playbooks that tell the model how each provision should be read.
• Apply Telon's validation methodology. Every extracted field is checked against the source clause, every discrepancy is logged with its reason, and accuracy is tracked at field level against the target.
• Work inside the client's Slack or Teams to a clear weekly plan and a short monthly status.
• Feed what you learn back into Telon's playbooks and the Outcomes Engine.

What you will bring

• A qualified lawyer with around three to five years in practice, including corporate, M&A or commercial contract work at a firm or in-house. You have reviewed the kinds of documents you will now be judging at scale: share purchase agreements, shareholder agreements, financing and security documents, and commercial contracts.
• The ability to open a definitive agreement, find the provisions that matter in a transfer or financing, and form a confident, defensible view on each.
• A clear sense of where interpretation ends and a genuinely novel tax or regulatory question begins.
• Precision and discipline. You are comfortable working to a defined validation method, logging discrepancies honestly and being measured against a field-level accuracy target.
• Real comfort using AI tools for structured extraction. You do not need to be a software engineer, but you should already be using models to do legal work and judging their output.
• Fluent professional English, written and spoken, and the ability to work US Eastern hours.
• The ability to run your own workstream inside a client team while staying closely connected to the Telon team.

Nice to have

• Cross-border or international transactional experience, particularly for US clients.

• Hands-on prompt engineering or structured extraction with tools such as Claude or ChatGPT.

• Experience with legal AI or contract lifecycle platforms.

• Excellent English Proficiency

Working pattern

Remote from Buenos Aires or elsewhere in Latin America, with full overlap with US Eastern hours.