Viacheslav (Slava) Rudenko and Julia Neusner - CodeX Group Meeting May 21, 2026

Why It Matters

By automating complex visa documentation while preserving data privacy, the tool could dramatically reduce legal costs and broaden access to U.S. immigration pathways for skilled professionals.

Key Takeaways

• Open-source AI tool streamlines O‑1A visa petition preparation.
• Integrates Claude LLM with custom USCIS decision‑flow knowledge base.
• Runs locally, keeping applicant data private and off SaaS servers.
• Provides risk flags, success probability, and lawyer‑review prompts.
• Seeks community partners to validate, improve, and scale the solution.

Summary

The CodeX group meeting featured Vyacheslav “Slava” Rudenko unveiling an open‑source AI application designed to automate the preparation of O‑1A immigration petitions. Drawing on his personal refugee experience, Rudenko built a workflow that couples Claude’s large‑language‑model capabilities with a curated knowledge base of publicly available USCIS case precedents and eight diagnostic criteria.

The tool conducts a nine‑phase intake, matches the petitioner’s profile against similar cases, and generates a complete petition draft in roughly fifteen minutes. It flags red‑flag items, highlights green‑flag evidence, and presents a success probability, while explicitly prompting users to have a qualified immigration attorney review the output. By eliminating the costly intermediary agency, the system aims to cut attorney fees that typically range from $6,000 to $12,000.

Rudenko emphasized that the solution is not a replacement for lawyers but a collaborative aid: “It saves money and time, and it tells you when the case is not viable.” He demonstrated a real‑world example involving a rejected petition for a former product manager, showing how the system identified procedural errors. The software runs locally from a GitHub repository, ensuring that user data never leaves the applicant’s computer.

If adopted widely, this approach could democratize access to high‑quality immigration counsel, lower barriers for skilled migrants, and spur community‑driven enhancements. Legal‑tech firms may see a new open‑source foundation for building cost‑effective immigration services, while regulators will monitor the balance between automation and professional oversight.

Original Description

CodeX Group Meeting | May 21, 2026

In this session of the Stanford CodeX group, we hear from two legal innovators using AI to tackle real-world problems.

Viacheslav (Slava) Rudenko — an AI product manager and Ukrainian refugee — shares his Claude-powered tool for O-1A immigration petitions. Built from public AAO case precedents and insights from retired USCIS officers, the tool walks users through all eight petition criteria, flags red flags, and generates a structured case package ready for attorney review — potentially cutting $10K in agency fees. He discusses his open-source aspirations and the guardrails he built in to prevent hallucinated citations.

Julia Neusner — Stanford Law '20 and director of the Human Impact Project — demos her AI-assisted tracker of immigration enforcement incidents across the U.S. Built with Claude Code by a self-described non-technical researcher, the tool scrapes news, litigation (via Court Listener), GoFundMe, and social media to document detainments, deportations, use-of-force incidents, and more. With nearly 2,500 geolocated incidents and growing, it's designed to support advocates, researchers, and attorneys working on habeas petitions, class actions, and policy challenges.