I'm a hands-on technology leader and founding engineer with 20+ years building production platforms, from Generative AI and B2B SaaS to connected mobility, IoT, cloud infrastructure and enterprise integrations.
I combine deep technical execution with strategic advisory: designing resilient systems, leading cross-functional engineering teams, and aligning technology with business goals from concept to production. My work spans three recurring themes:
“My experience across a broad, diverse range of technologies and platforms has given me the ability to learn, adapt, and build new technologies quickly and intelligently.”
Building production LLM features, agentic workflows, RAG and tool-calling, into real SaaS products.
Building software end-to-end, microservices, REST APIs and React front-ends, across multiple languages.
Modeling and moving data at scale, relational, NoSQL and streaming, behind clean, well-integrated APIs.
Architecting cloud-native infrastructure as code, containers, Kubernetes and multi-cloud deployments.
Automating delivery and keeping systems observable and reliable, CI/CD, monitoring and incident response.
Connecting vehicles, sensors and edge devices to the cloud through real-time telemetry.
Securing systems and traffic end-to-end, Zero Trust, encryption and SecOps automation.
Leading cross-functional R&D and turning innovation strategy into shipped products and proofs-of-concept.
Multi-tenant B2B SaaS that automates security-ticket triage, investigation and resolution with agentic LLM workflows, RAG and 40+ enterprise integrations.
Real-time fleet telemetry and connected-vehicle data platforms integrating OEM data layers and IoT infrastructure, delivered with OEM partners across three continents.
Sensor-driven accessibility interfaces using infrared, cameras and microcontrollers, pioneering low-cost inclusive keyboard access from existing hardware.
Secure, scalable automation-as-a-service frameworks and lab infrastructure for testing anti-phishing and malware-protection systems, built with a global team.
A reusable, retrieval-augmented chat widget: a hand-written FAQ layer in front of a knowledge-base-grounded LLM with prompt caching and a server-side safety filter. The same architecture that powers Avi'sBot on this site.
View sourceA double-opt-in contact form built for spam resistance: single-use SHA-256 magic-link tokens, App Check, per-IP and per-email rate limiting and a honeypot, the same flow behind the contact form on this site.
View sourceHome Assistant on Raspberry Pi with connected devices, an Arduino + AWS IoT Core relay project, and reverse-engineering a Domex smart-home controller.
Delivered a public lecture on Smart Mobility and Smart Cities as part of the Israeli Teachers' Training Program on sustainability.
Watch the talkThat chat bubble in the corner is Avi'sBot, a small, grounded AI assistant Avi built for this site. Ask it about his experience, skills or projects and it answers from a curated knowledge base rather than the open web, so replies stay accurate and on-message. It's a working demo of the same agentic-AI engineering Avi does day to day.
Answers visitor questions about Avi's background, experience, skills and selected work in plain language, and points you to the contact form for anything outside what it knows.
Common questions are answered instantly from a hand-written FAQ. Anything else goes to a grounded LLM that may only draw on Avi's portfolio as its source of truth, with a server-side safety filter on every reply, retrieval-augmented generation, end to end.
A serverless backend on Firebase Cloud Functions calls Claude (Anthropic) with prompt caching for speed and low cost, behind App Check and per-visitor rate limiting, the same production stack Avi uses in his work.
We sent a confirmation link to . Click it to deliver your message, it won't reach Avi until you do.
The link expires in 15 minutes. You can close this window; nothing is lost.
That confirmation link timed out. Send a fresh one and check your inbox again.