Michael Vexler Senior Full-Stack & Applied AI Engineer Agents, RAG & Evals in Production

Ramat Gan, Israel 32.08°N 34.81°E

a non-biological organism living in inner-product spaces

01 — how it actually works underneath 02 — what I build on top of it

Co-Founder & CTO of an AI-native live-events platform. Senior front-end before that — banking, payments, payroll, HR, cybersecurity.

Open to senior full-stack / applied-AI roles

michaelvx@gmail.com

01 — Production

Things shipped to production

one request: out of the tree, through the boundary, down the layers, and back

01 — the client: one session, a tree, its state 02 — the server: layers, shared and concurrent

One request: out of the component tree, through the validating boundary, down the server layers, and back

  1. AI platform 2026 — Present

    ShowMe

    Co-Founder & CTO

    A demand-driven live-events platform, in production. I own the architecture end to end, including the LLM agent that runs inside it.

    • Production LLM agent across web and the WhatsApp Business API: a provider-agnostic LLM gateway with failover, RAG over the live catalog, tool calling, and structured actions with human handoff.
    • Durable agent engine — queued multi-step ReAct loops, real-time streaming, loop detection, budget caps.
    • LLM evals & guardrails — LLM-as-judge scoring, golden regression gates in CI, jailbreak-resistant sanitization, calibration tracking.
    • Personalization — an embedding pipeline with per-user taste vectors on off-the-shelf embedding models.
    • Sababot — a WhatsApp-for-business product, now a major module inside ShowMe, with evals of its own.
    • Tooling — a layer over Claude Code's harness, used day to day.
    • React 19
    • TypeScript
    • Supabase (Postgres + RLS)
    • Multi-tenant RBAC
    • 120+ production migrations

02 — Lab

Own products, end to end


  1. SaaS 2024 — Present

    OsekOS

    Founder / Full-Stack Engineer

    A WhatsApp-first, multi-tenant SaaS on the WhatsApp Business API.

    • Authentication, billing, real-time messaging, production infrastructure.
    • Built with AI-assisted development — Claude, Gemini, and MCP tooling in the loop.
    • Angular 21
    • NestJS
    • GraphQL
    • Prisma
    • WhatsApp Business API
    No public link
  2. AI video analytics 2024 — Present

    FootAnalytics — MatchCut

    Founder / Full-Stack Engineer

    AI video analytics built on event-driven microservices. I worked on the ML inference pipeline and GPU acceleration — integration and cost optimization, not model training.

    • Event-driven microservices feeding an ML inference pipeline.
    • GPU acceleration — integration and cost optimization across the pipeline.
    • Event-driven microservices
    • ML inference pipeline
    • GPU acceleration
    No public link

03 — About

How I work


I build the layer above the model — where training and fine-tuning end, and agents, retrieval and evaluation begin.

Co-Founder & CTO of an AI-native live-events platform: an LLM agent, RAG, and an evaluation harness, in production.

Before that, senior front-end — banking, payments, payroll, HR, cybersecurity. Regulated environments, sensitive data, Angular at scale.

TypeScript full-stack (React, Angular, NestJS, Node.js), with Python for ML-pipeline integration.

AI end to end: agentic workflows, prompt & context engineering, MCP tooling, evals & guardrails.

04 — Experience

From front end to full stack

  1. 2026 — Present

    Co-Founder & CTO · ShowMe

    Demand-driven live-events platform, in production; owns architecture end to end.

    • Production LLM agent (web + WhatsApp Business API): provider-agnostic LLM gateway with failover, RAG over live catalog, tool calling, structured actions with human handoff.
    • Durable agent engine — queued multi-step ReAct loops, real-time streaming, loop detection, budget caps.
    • LLM evals & guardrails — LLM-as-judge scoring, golden regression gates in CI, jailbreak-resistant sanitization, calibration tracking.
    • Personalization — embedding pipeline + per-user taste vectors on off-the-shelf embedding models.
    • React 19
    • TypeScript
    • Supabase (Postgres + RLS)
    • Multi-tenant RBAC
    • 120+ production migrations
  2. 2024 — 2025

    Senior Front-End Engineer (Contract)

    • Discount Bank (2025) — enterprise Angular apps under banking security and compliance.
    • Spikerz (2024) — built & stabilized core product flows at an early-stage cybersecurity startup.
  3. Sep 2022 — Jan 2024

    Senior Front-End Engineer · E-Square

    Subcontracted to HiBob (HR SaaS) and Guardicore (cyber). Owned front-end architecture.

    • Front-end architecture: Nx monorepo, Storybook, Akita.
    • Testing pipelines (Cypress); AWS + GitHub Actions CI/CD.
  4. 2021 — 2022

    Senior Front-End Engineer · Opal Payment

    Fintech — high-reliability payment flows.

    • Angular
    • Azure
  5. 2020

    Senior Front-End Developer · Malam Payroll

    Payroll at national scale. Front-end architecture, coding standards, mentoring.

  6. 2015 — 2020

    Front-End Engineer

    Banking, FX trading and financial products.

    • Bank Leumi (2019 — 2020), via Matrix — front-end work inside the bank.
    • TradAir (2017 — 2019) — Angular, legacy migrations, UI performance.
    • TechFinancials (2015 — 2017) — HTML/CSS/JavaScript: landing & marketing pages.

Education

  • SCE — Shamoon College of Engineering B.Sc. Electrical & Electronic Engineering · 2009 — 2013
  • John Bryce Java EE Full-Stack · 2014 — 2015

Languages

  • Hebrew (native)
  • English
  • Russian

05 — Skills

The toolkit

AI / LLM

  • LLM integration & orchestration
  • RAG
  • Embeddings
  • Vector search
  • Agentic workflows
  • Tool calling
  • MCP
  • Prompt & context engineering
  • LLM evaluation & LLM-as-judge
  • Guardrails & safety filtering
  • Cost & latency optimization

Languages & Testing

  • TypeScript
  • JavaScript
  • Python (working — ML pipeline integration)
  • Cypress
  • Jasmine
  • Karma
  • TDD

Back-End & Cloud

  • NestJS
  • Node.js
  • GraphQL
  • REST
  • WebSockets
  • Prisma
  • PostgreSQL
  • Supabase
  • Serverless & edge functions
  • AWS
  • Azure
  • Docker
  • GitHub Actions CI/CD
  • Microservices
  • Event-driven architecture

Front-End

  • React 19
  • Angular (2–21)
  • RxJS
  • Signals
  • NgRx
  • Akita
  • shadcn/ui
  • Storybook
  • Nx monorepo