Professional profile

Legal engineering,
knowledge management
and AI governance

I am a Senior Attorney and Legal AI Operator admitted in New York and Brazil, with 15+ years of legal experience spanning practice and legal innovation. My practice experience started with litigation and transactions, and currently focuses on regulatory and corporate matters. Over the years, I have worked in legal engineering, knowledge management and artificial intelligence (AI) governance. I have led legal teams for the past decade, an 8-person multi-tier team in Brazil and a 5-person multi-tier legal team across 150 concurrent matters in the U.S.

Alongside the practice, I build the systems legal work runs on: workflows taken apart by the kind of work each step performs, AI tools grounded in sources a lawyer can open, and controls that keep release with the responsible attorney. My articles and case studies apply that method to conflicts intake, knowledge management, prompt engineering and AI governance.

Scroll to explore the work

Selected work / Evidence

The work, up close.

Articles, an illustrative case study, an application analysis and a configurable demonstration.

Knowledge management article

The Modern Precedent Lifecycle Workflow

A proposed workflow for capturing attorney reasoning, structuring source-linked resources, testing AI grounded in those resources and delivering it where lawyers draft. The article connects knowledge management with adoption and the economics of recurring legal work, including an 8-category pre-adoption test harness.

Read article : The Modern Precedent Lifecycle Workflow (opens online)Next work
Illustration of a proposed workflow04 / 04

Drafting workspace

Original document → reusable record → source-linked guidance. The responsible lawyer reviews the work.

Use

Source-linked guidance reaches the drafting workspace. The responsible lawyer retains the final say.

A completed document becomes reusable, source-linked knowledge. Testing precedes use; the responsible lawyer keeps the decision.

Scroll to follow the explanation

Legal engineering case study

Workflow Automation: Conflicts

An illustrative 9-step conflicts and new-matter intake design that assigns drafting, source-backed retrieval, fixed rules and attorney judgment to the right parts of the workflow. The design places access checks before retrieval, preserves the sources behind the work and keeps adversity, clearance and engagement-letter release with counsel.

Read case study : Workflow Automation: Conflicts (opens online)Next work
Illustrative case study · nine-step workflow03 / 03

Counsel owns the decision

Steps 6 and 7: assess adversity and decide clearance. Step 8: the engagement letter waits for attorney release.

Counsel

The attorney assesses adversity and decides clearance. The engagement letter also waits for authorized release.

An intake request expands into a nine-step workflow. Each step is allocated by work type. Counsel retains adversity and clearance decisions and release of the engagement letter.

Scroll to follow the explanation

Application analysis

Prompt Engineering (O’Mono)

6parallel worktrees
1,426test assertions
0failures in that build

Built across desktop and web, with task interpretation, versioned model profiles, prompt compilation, local privacy screening, human editing and validation. A documented build in Claude Code used 6 parallel development worktrees and recorded 1,426 test assertions with 0 failures in that build.

Read application analysis : Prompt Engineering (O’Mono) (opens online)Next work
Curriculum concept · August 2026 build specification04 / 04

Curriculum shaped by work

The specified curriculum links recurring risks to targeted training and distinguishes learning needs from system fixes.

  1. Recognize task-specific risks
  2. Understand AI and its limits
  3. Review and verify outputs
  4. Instruct AI effectively
  5. Choose systems and uses
  6. Retain human responsibility

Curriculum

The user’s exported learning evidence shapes targeted modules and short lessons. System and workflow fixes remain separate.

Curriculum concept from the August 2026 build specification: task-specific lessons feed local learning evidence; the user exports that evidence manually for curriculum development. The current Mac implementation remains unverified.

Scroll to follow the explanation

Configurable governance framework

AI Governance Operating Model

A configurable operating model connecting AI use with responsibility: a model and system registry, risk classification, vendor diligence and escalation. It generates organization-specific governance documentation, including framework crosswalks, incident handling and principles, mapped to NIST AI RMF, ISO/IEC 42001, the EU AI Act and ABA professional-responsibility requirements.

View demonstration : AI Governance Operating Model (opens online)Continue to expertise
Illustrative operating arrangement03 / 03

An issue has somewhere to go

Escalation keeps the issue, assigned reviewer and response in view. The diagram illustrates an operating arrangement.

Escalation

An unresolved issue reaches an assigned reviewer through a defined escalation path. The record keeps the issue and responsibility visible.

Illustrative operating arrangement: the same AI system record stays connected to its owner, risk and vendor review responsibilities, and an escalation route. No live system behavior or compliance result is asserted.

Scroll to follow the explanation

Expertise

Expertise in detail

Legal engineering and workflow design

I map legal work across actors, inputs, handoffs and decisions, then separate drafting, source-backed retrieval, fixed-rule configuration and attorney judgment. The map becomes a scoped use case, technical requirements, controls and tests. I build legal AI systems with OpenAI and Anthropic application programming interfaces (APIs), Retrieval-Augmented Generation (RAG), which connects model output to retrieved source material, Model Context Protocol (MCP), which connects AI tools to data and services, Python, JavaScript, n8n, Claude Code and Codex.

Knowledge management and attorney resources

I convert recurring attorney reasoning into model documents, legal memoranda, playbooks, decision rules and standard operating procedures, and connect those assets to their sources, intended use and review history. My work includes memoranda on novel legal theories that partners and associates adopted as practice templates, internal-resource retrieval and AI workflows grounded in the documents lawyers already use.

AI governance, implementation and adoption

I design controls for data handling, risk classification, vendor review, testing, human oversight and escalation, mapping governance documentation to the National Institute of Standards and Technology AI Risk Management Framework (NIST AI RMF), ISO/IEC 42001, the international standard for AI management systems, the European Union AI Act and American Bar Association (ABA) professional-responsibility requirements, including Formal Opinion 512. I carry adoption through workflow integration, training and feedback; my experience includes a 2021 practice-wide platform rollout and a compliance-screening pilot that reached first value in under 8 weeks and utilization above 80% at 6 months.

Professional experience

The complete record.

February 2015 to Present

Legal Innovation Arc | jhmacal.com

Legal Engineering, Knowledge Systems and AI Governance

A continuing body of legal innovation work, beginning with law-firm digital transformation and developing into application engineering, knowledge systems, AI governance and professional education alongside legal practice.

  • Map legal work across actors, inputs, handoffs, decisions, judgment boundaries and data constraints, then define use cases, controls, pilot criteria and implementation requirements; published an illustrative conflicts and new-matter intake case study separating a 9-step workflow into drafting, source-backed retrieval, fixed-rule configuration and attorney judgment.
  • Built and maintain O’Mono across desktop and web, interpreting tasks, compiling instructions for selected AI systems through versioned capability profiles, and embedding validation, local privacy screening, human editing and structured outcome evidence; rebuilt a documented version in Claude Code across 6 parallel development worktrees, with 1,426 test assertions and 0 failures in that build.
  • Engineered a configurable AI governance operating model that generates organization-specific documentation, with intake and routing logic, a model and system registry, risk classification and annexes for vendor diligence, framework crosswalks, escalation, incidents and principles, mapped to NIST AI RMF, ISO/IEC 42001, the EU AI Act and ABA professional-responsibility requirements.
  • Designed an AI tool-control architecture with a 20-surface inventory, a vendor register recording training-on-inputs and retention terms with clause citations and review dates, a 32-item intake checklist with connector and automation addenda, and 8 information-sensitivity classes with a hard bar on secrets.
  • Specified an 8-category pre-adoption test harness covering hallucination, citation validity, quote fidelity, record fidelity, privilege leakage, jurisdiction accuracy, summarization accuracy and bias, with tool-intake and pre-filing verification gates keyed to ABA Formal Opinion 512, NIST AI RMF and ISO/IEC 42001.
  • Stress-tested governance architecture through OpenAI and Anthropic model review, including an 11-agent parallel audit and a 249-subagent whole-tree review; converted confirmed bypasses and checker defects into targeted repairs and regression tests.
  • Built a provenance-rich public-data research platform over 11.35 million filings across 19 fiscal years, reducing the current-case data approximately fivefold and cutting organization-level analysis from hours to seconds in the September 2026 reviewed build, with source fingerprints, row-level provenance, revision history and quality flags.
  • Engineered document quality control covering fee and filing-address verification, form-edition checks, source-form fingerprints, exhibit numbering across conditional variants and render verification; keep legal judgment, evidence conflicts, signatures, payment details and final release with the responsible professional.
  • Create ground-truth datasets, grading specifications and scoring rubrics as a senior legal expert selected by frontier AI labs to benchmark and improve model performance.
  • Taught responsible AI adoption to MBA and graduate professionals at the McColl School of Business, Queens University of Charlotte, delivered a paid masterclass to 300+ attendees and authored responsible-adoption course materials; publish practical guidance including The Modern Precedent Lifecycle Workflow, Prompt Engineering (O’Mono), Legal Reasoning as a Strategic Asset and The AI Act After the Omnibus.

August 2025 to Present

Reinhardt Savic Foley LLP

Senior Associate

Practice leadership, attorney workflow integration, quality review and legal-technical delivery across an international matter portfolio.

  • Lead a 5-person multi-tier legal team and a portfolio of 150 concurrent matters across North America, Latin America and Europe, the Middle East and Africa (EMEA) for clients in heavily regulated industries such as financial services, technology and fintech, directing intake, allocation, prioritization, review, quality control, benchmarks and escalation.
  • Develop the team through onboarding, professional development, formal evaluations, compensation recommendations and shared hiring decisions, with defined review responsibilities and benchmarks for legal delivery.
  • Advise employers and business professionals on employment terms, worker classification, payroll practices, remote-work policies and terminations; conduct privileged compliance audits and remediation, set records-retention schedules and train human resources (HR) teams.
  • Represent clients in federal administrative appeals and motions, defend employers in Department of Labor audits, wage investigations and inspections, review evidentiary responses to federal-agency audits before filing, and submit Freedom of Information Act (FOIA) requests, rulemaking comments and regulatory advisories.
  • Map attorney work as actually performed, separate the judgment that must stay human from work that can be standardized or automated, and build standards and tools for the rest, including agent-assisted retrieval across internal resources.
  • Ran an n8n pipeline that collected agency alerts, newsletters and primary sources from 50+ feeds across jurisdictions, removed duplicates, ranked primary sources first, delivered a Monday brief and escalated flagged items through a rubric; later moved the same collection and triage to Codex and Claude scheduled tasks.
  • Set standards for structured outputs, source traceability, human review, validation, privilege, confidentiality and escalation in AI-assisted legal work, and translate them into designs, tests and implementation decisions with technology, information security, engineering and analytics colleagues.

October 2020 to August 2025

Reinhardt Savic Foley LLP

Associate

Corporate, transactional and regulatory practice, reusable practice resources and AI adoption inside attorneys’ existing workflows.

  • Represented business owners in the purchase and sale of companies across 3+ acquisitions, including a software startup’s sale to a major technology company; negotiated deal terms, drafted letters of intent, acquisition agreements, disclosure schedules, transition services agreements and retention agreements, conducted due diligence and managed post-closing integration.
  • Architected corporate and governance structures for domestic and foreign businesspersons, mapping ownership and control, and prepared capitalization records, board resolutions, bylaws, operating agreements and director agreements to allocate authority and satisfy statutory and regulatory requirements.
  • Drafted, reviewed and negotiated customer, vendor, contractor, staffing, client services and intercompany agreements, along with publishing and agency agreements, event deal memos, commercial leases and lease assignments, and reviewed franchise agreements and disclosure documents.
  • Led evaluation, selection and the 2021 practice-wide rollout of a legal AI workflow platform, establishing operating standards, adoption practices and user-feedback loops that remain in use.
  • Identified manual compliance screening against scattered client records, designed and piloted a replacement built on a large language model (LLM) with Retrieval-Augmented Generation, blind-tested its outputs against historical senior-partner assessments and integrated it into attorneys’ existing tools; the pilot reached first value in under 8 weeks and utilization above 80% at 6 months.
  • Prepared evidentiary responses in federal-agency audits, about 30 a year at the outset, agency audit responses, about 5 to 10 a year, and administrative appeals of agency decisions, with responsibility for source records and the reasoning supporting the submission.
  • Authored model documents and legal memoranda, including memoranda on novel legal theories adopted by partners and associates as practice templates, and developed standard operating procedures for document review, deadlines and client communication.
  • Managed major client accounts directly, translated regulatory and compliance requirements into executive-ready guidance and built a standing compliance offering for corporate clients around audit readiness, documentation controls, records governance and federal enforcement response.

July 2020 to October 2020

Columbus Consulting Group

Associate

Matter coordination and client guidance in New York.

  • Managed operational workflows and high-volume portfolios, advising executives and multinational organizations on legal strategy, scheduling, technical criteria and timelines.

December 2017 to August 2019

Zamenhof Gurgel Macedo

Senior Attorney

Team leadership, complex litigation and negotiation in Brazil.

  • Litigated tax controversy, public-law and government-contract disputes, constitutional cases, post-judgment enforcement and habeas corpus appeals for corporate and public-sector clients as counsel of record and drafting attorney, through state courts, the federal courts of the 5th Region, electoral courts, the Superior Court of Justice (STJ) and the Supreme Federal Court (STF).
  • Ran 50+ active cases a month and argued hearings in roughly half of them; drafted pleadings challenging tax assessments before and during enforcement and opposing enforcement of judgments, and built the evidentiary defense that extinguished a tax assessment above 2.2 million Brazilian reais (BRL) for a public-sector client.
  • Authored actions that released federal transfers withheld from municipalities, public-integrity civil actions against former public administrators for losses to public entities, and a direct constitutional challenge to a state audit court’s rule, filed in the name of 50 mayors on a thesis courts had already adopted in earlier suits.
  • As of June 2021, named in 2,700+ published decisions, including 300+ in Brazil’s superior courts, with decisions adopting theses I authored.
  • Led the legal team for 4 state campaigns in the 2018 election cycle, covering candidacy registration, campaign-finance filings and complaints against opposing campaigns under rapidly changing rules.
  • Directed an 8-person multi-tier team and implemented quality-assurance protocols, key performance indicators (KPIs) and review mechanisms for high-volume legal delivery.
  • Led negotiation and execution of a cross-border brand-licensing agreement between a Brazilian beverage manufacturer and a U.S. consumer brand, from term design through closing, securing the license to produce and use the brand in Brazil, and performed exact and similar-mark trademark-clearance research.

February 2015 to November 2017

Zamenhof Gurgel Macedo

Attorney

Litigation, election-law counsel and law-firm digital transformation in Brazil.

  • Represented corporate and public-sector clients in tax, administrative, commercial, employment, contract and election-law matters as counsel of record and drafting attorney, from initial pleadings and evidentiary filings through appeals.
  • Led the legal team for 8 municipal campaigns in the 2016 election cycle, providing counsel, regulatory support and legal strategy under rapidly changing rules.
  • Led software vendor selection, hardware modernization, a centralized knowledge library and a unified legal system as the firm tripled revenue and headcount.

February 2013 to January 2015

Zamenhof Gurgel Macedo

Paralegal

Foundational litigation research and drafting experience.

  • Supported the firm’s litigation matters through legal research and drafting.

January 2011 to January 2013

Courts of Law

Clerk, Intern and Volunteer

Court experience in Natal, Brazil.

  • Conducted legal research, assisted judges with draft orders and decisions, attended hearings and trials, and prepared notes that informed decisions.

Education and credentials

Education, admissions
and recognition

Education

Fordham University School of Law

Master of Laws (LL.M.) in Intellectual Property and Information Technology Law

May 2020

Merit-Based Scholarship Recipient

Federal University of Rio Grande do Norte

Bachelor of Laws (Juris Doctor equivalent)

January 2015

Admissions

  • New York State Bar
  • Brazilian Bar Association (Ordem dos Advogados do Brasil, OAB)

Languages

  • English: Fluent
  • Portuguese: Fluent
  • Spanish: Proficient

Recognition

Founding Member, Reframe.Lawyer
AI in Legal Practice

EB-1A Extraordinary Ability Recipient
U.S. Permanent Resident; no sponsorship required

Guest Lecturer, McColl School of Business, Queens University of Charlotte
Responsible AI adoption for MBA and graduate professionals