New York · Admitted NY & Brazil
Senior Attorney & Legal AI Operator. Workflows rebuilt on AI and integrated where lawyers work. Systems built to scale.
Senior attorney in New York, specialized in IP and IT Law, designing legal workflows on AI.
Decomposes workflows, designs the steps that should be rebuilt on retrieval and language models, and integrates the result into the tools, driving adoption through ordinary use. Embeds the controls underneath so the speed survives review and earns attorneys trust, optimizing enablement.
Roles
Takes a group's working process apart step by step, separating what a model should do from what belongs to deterministic logic and what stays with a lawyer. Rebuilds the parts worth rebuilding and leaves the rest alone.
Builds AI tooling end to end. Custom research pipelines, agentic systems, retrieval design, and the data infrastructure behind them. Practitioner-built, on real work.
ROI lives in depth of usage. Prioritizes systems inside the surfaces attorneys already open rather than behind another login.
Designs the controls, decision records, and release gates that let a firm move quickly without losing the audit trail. Governance built to keep adoption defensible instead of slowing it down.
Principles
Operates inside the working cycle rather than reviewing it afterward. Present when the process is designed, when the system is chosen, and when the record has to support the decision. Legal judgment moves into the build.
Builds end-to-end legal programs designed to absorb growth, regulatory shifts, and team transitions without breaking. Workflow architecture, operating protocols, and the cross-functional structure that holds them together. Hires, develops, and manages the team that runs them.
Engineers AI-assisted workflows for intake, research, classification, reporting, and quality control, with retrieval designed so the system can be right about the things that matter. Privilege-safe by design, and instrumented so usage can be measured rather than assumed.
Translates dense regulatory and technical complexity into decisions that executives, lawyers, operators, and engineers can act on. The communication layer is part of the work. Adoption follows understanding, and understanding is something you design for.
Scope
Maps how a legal team actually works, including the exceptions that consume most of the week, then rebuilds the process on a substrate that fits it. Separates the steps a model should handle from the steps that require deterministic logic or professional judgment.
Designs and builds the systems themselves. Retrieval architecture, index composition, source authority hierarchies, staleness handling, agentic pipelines, and the data infrastructure underneath. Built to be inspected, corrected, and maintained by the people who inherit it.
Gets systems used. Integration into existing tools, instrumentation that shows whether usage is real, and the reviewer guidance that matters once model output enters the workflow. Adoption designed into the system rather than delivered as training.
Owns the control layer. Decision records, release gates, escalation paths, and documentation held at audit grade. Governance positioned as the thing that makes fast adoption defensible to a regulator, a client, or a court.