$ About Me
I'm a senior software engineer specializing in AI/LLM systems and full-stack delivery. I've been building production AI since 2019, including one of fintech's earliest LLM deployments, and I've since gone deep across the full applied AI stack: agentic pipelines, ML inference optimization, LLMOps, vector search, and security-hardened LLM middleware.
My career started at Intuit, where I shipped a Python LLM chatbot for QuickBooks and built the AI agent platform that cut CX support costs by 40% across QuickBooks and TurboTax. From there I led authentication infrastructure at EarnIn, deploying OAuth 2.0 and Passkey auth to millions of users. Most recently I've been building pay recommendation AI and an LLM compensation agent at Syndio.
When I'm not coding, you can find me exploring new technologies, reading books on science and history, or sharing knowledge through technical writing and mentoring.

Josh Janes
Senior Software Engineer · AI / LLM
// Technical Skills
AI / LLM
Languages
Backend & Cloud
DevOps & Observability
Frontend & Mobile
// Experience
AI Software Engineer
Syndio · Remote
Building pay recommendation ML API end-to-end: REST API design, async Pub/Sub + Cloud Functions inference pipeline, and multi-model Vertex AI predictions from a single endpoint, achieving 2-4x inference speedup. Migrated Vertex AI Vector Search to pgvector, eliminating managed vector service costs with no retrieval quality loss. Built PII redaction middleware using Microsoft Presidio + spaCy NER to satisfy enterprise compliance requirements. Architected an AI compensation agent with Google ADK and Vertex AI Gemini.
Indie Developer & AI Research
Self-directed
Shipped Slide 2.0, a full Android game, to the Google Play Store. Independently researched multi-agent AI systems and agentic frameworks.
View on Google Play →Backend Software Engineer
EarnIn · Remote (US)
Led a new authentication platform (TypeScript/React frontend and C#/.NET microservices backend) deployed to millions of users across mobile and web. Implemented OAuth 2.0 and Passkey authentication, reducing security vulnerabilities by 15% by replacing a fragile legacy auth architecture. Keycloak integration reduced customer support call volume by 20%. Conducted technical interviews for IC Software Engineer candidates.
Independent Study
Self-directed
Explored early GPT-3-era language modeling and applied NLP through personal projects ahead of the generative AI wave.
Software Engineer II
Intuit · Mountain View, CA
Deployed a Python LLM chatbot for QuickBooks Assistant, one of fintech's earliest production LLM deployments, driving +14% message engagement. Built the Java/Spring Boot AI agent platform for QuickBooks and TurboTax, automating customer query resolution and delivering a 40% reduction in CX support calls. Migrated from synchronous REST to asynchronous gRPC, enabling real-time chat at scale. Promoted L3 to L4 within 13 months.
Software Developer Intern
Intuit · Edmonton, AB
Rebuilt the first-time user experience for QuickBooks Online, refactoring legacy JavaScript into a modular React SPA validated with A/B testing.
// Let's Connect
I'm always interested in hearing about new projects and opportunities. Whether you have a question or just want to say hi, feel free to reach out!