Tony Virgil

Tony Virgil

19 | Builder & Engineer | Voice AI & Dev Tools

Former electronics engineering student, 19. Shipping AI products since Feb 2025. Pattern: build fast, ship faster, learn from failures, repeat. Reached $500k+ AgTech pipeline in <9 months. Currently building Voice AI for real estate & memory systems for AI agents.
Now Building

Casius — Voice AI

Problem: Dubai agents waste 6+ hours daily on repetitive buyer questions.

Solution: AI answers in <60s, speaks 4 languages, qualifies leads in real-time.

Stack: Next.js, Node/Python, MySQL, VPS. Custom voice pipeline + Meta integrations.

Status: MVP shipped. First agency in pipeline.

Response
<60s
Languages
4
Time to MVP
Weeks

Grov — AI Memory

What: Captures reasoning from Claude Code sessions, injects context into future ones.

Impact: 1-2min execution vs 10+min explore. <2% tokens vs 7%+.

Stack: SQLite, GPT-3.5 extraction, semantic filtering, local-first.

Status: v0.1 ready. Launching on npm/GitHub/HN soon.

Exploration
1-2m
Token Save
-85%
Build Time
48h
Past Work

SSA — AgTech Hardware + ML

Achievement: $500k+ pipeline, 54% margins in <9 months. Enterprise pilots across 3 countries.

Built: IoT + neural networks for crop automation. Full-stack: Swift app, PyTorch models, hardware design.

Lesson: Validate faster. Sell before building.

EF Builder's Retreat (Zurich)

Achievement: 1 of 23 selected (100+ applicants). Youngest participant. Peer selection winner. Also got cool stickers and merch.

Project: Browser extension: SMB owners modify websites via plain English → push to production.

Base Jump Arena (Hackathon)

Achievement: 1/7 Romanian selected, 1/30 total for weekend-long hackathon.

Built: Dimini — AI therapy assistant with real-time semantic relationship visualization using GPT-4 + embeddings.

Technical Stack
Frontend: React, Next.js, Astro Backend: Node, Python, FastAPI ML: PyTorch, custom training Infra: VPS, Docker, databases Foundation: Raised on C/C++, traded memory management for velocity