Joseph Blas
AI Developer · Autonomous Systems Builder · Multi-Agent AI Systems Builder
Self-taught engineer — 10 years building autonomous vehicle systems, then 5+ years learning AI tools, code, and shipping personal and client projects. No degree, all hands-on. Started at Google's Self-Driving Car Project (now Waymo) in 2009, where I built the Firefly autonomous vehicle from the ground up, became the drive-by-wire SME across 100 vehicles and 10 hardware iterations, and reduced sensor calibration from 9 hours to 60 minutes. I was part of the team that supported the world's first fully driverless ride on public roads in 2015.
After Google, I led system integration for Uber ATG's 10-truck autonomous fleet (~90% uptime), then joined Pronto.ai as the sole technician — building and operating the entire fleet for a 2,900-mile cross-country autonomous demo with zero critical failures.
Now I build production AI systems. Hermes is a multi-agent AI system I operate as CTO, with Lurkr (me) as CTO, executive skills (Chief of Staff, CFO, COO, CMO), VPs for Engineering, Infrastructure, and Product, and 14 agents. Main agent and sub-agents are routed to Ollama Cloud models by task — I pick the model that fits the job. It is open source at github.com/joestechsolutions/hermes-forge. I previously built OpenClaw, a 22-agent specialized system (2024–2026, now retired); the migration postmortem is on this site. The Skate Workshop is a React Native app for an Olympic-level skateboarding coach with 40+ Supabase tables and Stripe Connect. I also built this portfolio site (the AI chatbot is real, ask it anything), a DALL-E image generator, and Whisper Walkie (push-to-talk voice typing, 100% local processing).
Seeking AI Developer, Autonomous Systems, and Embedded/Robotics roles in the USA
Experience
Founder and Builder
Joe's Tech Solutions LLC — Hermes (multi-agent AI team: Lurkr as CTO, C-suite skills, VPs, 14 agents via Ollama Cloud model selection), React Native apps, full-stack AI development, open-source AI tooling
Autonomous Vehicle Technician
Pronto.ai — Sole technician — built entire fleet, 2,900-mile autonomous demo, zero critical failures
Autonomous Truck Technician
Uber ATG (Otto) — System integration for 10-truck fleet, ~90% uptime, drive-by-wire integration
Program Manager L4 / Fleet Vehicle Technician
Google Self-Driving Car Project (Waymo) — Built Firefly from ground up, drive-by-wire SME, calibration 9hr→60min, first driverless ride
Projects
Hermes
Multi-agent AI team: Lurkr as CTO, executive skills, VPs for Engineering/Infra/Product, 14 agents routed through Ollama Cloud model selection · open source at github.com/joestechsolutions/hermes-forge
Hermes: The OpenClaw Migration (case study)
Why I retired a 22-agent specialized system for a composable Hermes system — design for composability, not specialization
The Skate Workshop
React Native app for Olympic coach — 40+ Supabase tables, 18+ Edge Functions, Stripe Connect
cv-joseph
This portfolio — React 19, TypeScript, Tailwind, Vite, Claude API chat with RAG + voice mode
Career Ops
AI-powered job search pipeline — Claude Code, Go dashboard, ATS PDF generation, 45+ pre-configured portals
DALL-E Image Generator
Production AI image generation application
Whisper Walkie
Push-to-talk voice typing tool — 100% local processing
Certifications
AWS / Google
Meta
University / Academic
IBM / Cloud Academy
Education
- Self-taught — 10 years in AV programs + 5+ years learning AI tools, code, and building projects
- No formal degree — professional certifications in cloud, AI, platform engineering, and cybersecurity
Press
Pronto.ai — 2,900-Mile Autonomous Cross-Country Demo
Vimeo · 2019
Community
Frequently Asked Questions
Who is Joseph Blas?
Joseph Blas (also known as Joe Blas) is an AI Developer and Autonomous Vehicle Systems Veteran based in Escondido, San Diego, CA. He is self-taught with no formal degree: 10 years hands-on in autonomous vehicle programs (Google/Waymo, Uber ATG, Pronto.ai) and 5+ years learning AI tools, code, and building personal and client projects. He started at Google's Self-Driving Car Project (Waymo) in 2009, where he built the Firefly autonomous vehicle from the ground up, served as drive-by-wire SME across 100 vehicles, reduced sensor calibration from 9 hours to 60 minutes, and supported the world's first fully driverless ride on public roads in 2015. He went on to lead system integration at Uber ATG and was the sole technician at Pronto.ai, completing a 2,900-mile cross-country autonomous demo with zero critical failures. He now builds production AI systems through Joe's Tech Solutions LLC, operating Hermes as CTO through Ollama Cloud model selection. His full portfolio is at joestechsolutions.com.
What has Joseph built?
Joseph has built production systems across autonomous vehicles and AI. In AV, he built the Firefly self-driving car at Google/Waymo, integrated Uber ATG's 10-truck autonomous fleet, and built Pronto.ai's entire autonomous fleet. In AI/software, he built Hermes (multi-agent AI team: Lurkr as CTO, executive skills, VPs for Engineering/Infra/Product, 14 agents routed through Ollama Cloud model selection — open source), previously OpenClaw (22-agent specialized system, 2024–2026, now retired), Career Ops (Claude Code + Go dashboard + 45+ pre-configured job portals), The Skate Workshop (React Native + 40+ Supabase tables + Stripe Connect), a DALL-E image generator, Whisper Walkie (push-to-talk voice typing with 100% local processing), and cloud infrastructure with Terraform, Kubernetes, and Anthos/GKE.
What is Hermes?
Hermes is Joseph's current multi-agent AI system. Lurkr (himself) acts as CTO, supported by executive skills (Chief of Staff, CFO, COO, CMO), VPs for Engineering, Infrastructure, and Product, and 14 named agents. Main agent and sub-agents are routed to Ollama Cloud models by task — he picks the model that fits the job, not the brand. Agents compose via delegation rather than hard-wired specialization. It runs the real business operations of Joe's Tech Solutions: email, CRM, invoicing, SEO, deployments, content drafting, and 15 cron jobs. Skills system for reusable patterns. MemPalace for persistent memory across sessions. Open source at github.com/joestechsolutions/hermes-forge. The migration from OpenClaw (22 agents) to Hermes is documented in the OpenClaw case study on this site.
What certifications does Joseph have?
Joseph holds 12 professional certifications earned in 2024-2025. From AWS: Cloud Solutions Architect Professional. From Google: IT Support Professional and Cybersecurity Professional. From Meta: Programming with JavaScript, Version Control (Git), and Mobile Development. From University of Michigan: Programming for Everybody (Python). From Johns Hopkins University: The Unix Workbench (Linux). From IBM: Agile Development & Scrum and Cloud Computing. From Cloud Academy: AWS Cloud Bootcamp (59 hrs) and Azure Cloud Bootcamp (59 hrs). He has no formal degree — his expertise comes from 10 years in AV programs, 5+ years of learning AI tools and code, and continuous certification.