Building AI agents and
full-stack systems for
immigration law firms.

When Fireflies finishes an intake meeting transcript, this pipeline picks it up and hands it to an AI evaluator that extracts 14 structured fields: meeting outcome, retainer fee, visa fit, referral source, education, experience, location. Apollo fills in missing LinkedIn and company data, the transcript plus an AI summary PDF upload to HubSpot, and an optional Notion sync mirrors the case file. Runs as both a signed real-time webhook and a 5-minute HubSpot pipeline poll.

Long-form E-2 visa landing page for an immigration boutique. Design-to-code build: the client handed over a high-fidelity design and I shipped it pixel-accurate as a production Next.js + Tailwind + Framer Motion site.

An AI visa eligibility screener for Path Law Group. It reads contact questionnaire responses from HubSpot, runs real-time web research (LinkedIn, Google Scholar, press) to ground claims in verifiable facts, and evaluates the applicant against USCIS criteria for eight visa categories. The output is a structured verdict (qualified, borderline, weak) with a confidence score, criteria breakdown, and visa-specific narrative, written back to HubSpot as a formatted note plus a one-page attorney pre-meeting brief PDF. Calibrated against 977 historical Path Law cases.