Alon Yakoby
Available from December 2026 · open to full-stack & AI roles
Alon Yakoby, Full-Stack & AI Engineer

I build software that has to work on match day.

I’m a full-stack engineer with a backend focus in Swift, Python and TypeScript. I run the digital platforms of an Austrian football federation, build vision AI for a collector marketplace, and have shipped software for banks, insurers and a 200-room hotel.

  • Python & FastAPI
  • React & TypeScript
  • Vision AI
  • Sports Tech
  • CI/CD & Kubernetes
View selected work LinkedIn Contact
  • 8+ yrsbuilding and shipping software
  • 150matches scored live on my platform every Sunday
  • 8,000registered players on the federation platform
  • 80%+of the federation’s admin work automated
Vienna, Austria
~/ask-alonready

// systemConnected to my portfolio. Ask anything about my work, stack or availability.

Retrieval over this site’s content. How it works

Live in production: the ÖKFB platform

I designed, built and run the platform behind Austria’s small-field football federation: a public website, admin dashboard, team and referee portals and an app, on one Swift/Vapor backend with Stripe payments and AI match previews.

The ÖKFB public website with league selection, navigation and a news banner showing the team portal, a match card and an AI matchday report
An AI-generated German match preview on oekfb.eu with recent form for both teams
Two phones showing the ÖKFB app with a live match timeline and team statistics
  1. Public website: fixtures, live scores, tables, discipline and transfers
  2. Team portal: player profiles, registrations and photos
  3. Match cards: recent form, kick-off and venue for every fixture
  4. AI matchday reports: a report for every league
  5. AI match previews: written from facts the backend computes
  6. Mobile app: live ticker, pushes and team management
Screenshots from oekfb.eu, November 2026.Visit oekfb.eu ↗
  • Match day, liveReferees score around 150 matches every Sunday from their phones, straight into tables, leaderboards and suspensions.
  • Money you can trustStripe payments with verified webhooks, idempotent credits, versioned fees and automatic invoices.
  • AI that can’t invent resultsThe model only chooses among facts the backend has computed, and every article is reviewed before it’s published.
Read the case studyVisit oekfb.eu

Selected work

Things I’ve built

Products, platforms and experiments, from live sports software and payments to vision AI and native apps.

01/ 00

How I work

The same route runs through every project, from the first conversation with the people who will use it to running it in production. Select a stop to see what happens there.

1

Problem

Start from the people who will use it: referees entering a score on a phone, a receptionist checking in a guest, a collector listing a card. Write the problem down in their words.

How I work with AI

AI is something I put into products and a tool I use every day to build them. This is where I am moving my career, and here is how it shows up today.

AI in production

A model is only useful once it runs reliably behind an API.

  • Fact-grounded match previews
  • Structured JSON outputs
  • Durable AI job pipelines
  • Vision embeddings and vector search
  • Generated kit images

Building now

AI match analysis from video.

  • YOLOv8 on my own labelled data
  • Label Studio annotation
  • MCP with a vision stream model
  • Events, positions and player identity

AI-assisted engineering

Building faster without lowering the bar.

  • Claude Code
  • Prototyping
  • Test writing
  • Refactoring

Cost and reliability

Keeping AI features affordable and dependable.

  • AI and cloud cost control
  • Matching quality checks
  • Monitoring
  • Documentation

How the chat on this page works

  1. Your question
  2. BM25 retrieval over my knowledge base
  3. Top passages as context
  4. Model answers only from them
  5. Answer with sources

The knowledge base is written from this site, so the chat can’t claim anything the site doesn’t. Without a model configured, it answers directly from the best-matching passages. Same idea as my federation AI previews: compute or retrieve the facts first, let the model only phrase them.

For AI engineering in depth, see the AI match previews, vision matching at THATSIT! and the match analysis I’m building now.

This site is based on my wife Dafna’s portfolio, which I liked so much I adapted it for myself, with Claude Code as a pair programmer.

Toolbox

What I build with, across the whole stack. Pick a category to highlight it; the larger the word, the more I use it.

    Strongest skills

      Years since I started using each one professionally.

      Résumé

      8+ years building, shipping and running software, with growing ownership of architecture, releases and product decisions. Select a role for the details.

      THATSIT!

      Co-founder · Technical Product Lead

      Nov 2025 – PresentFounder

      Collector marketplace with over 400,000 cards and 2 million listings, launching in December 2026.

      • Built AI card matching in Python with FastAPI, PyTorch and Transformers: DINOv2 image embeddings and Qdrant vector search link each listing to the right card.
      • Run it as containerised services next to a React Native app and a Swift backend.
      • Own product vision, roadmap and launch plan; lead sprint planning and retrospectives.
      • Keep the budget in check, including the running costs of AI tools, cloud services and infrastructure.
      • Python
      • FastAPI
      • PyTorch
      • DINOv2
      • Qdrant
      • Docker
      • React Native
      • Swift
      Case study ↗
      Open printable résumé One page, with links to every project. Print it or save it as a PDF.

      Education

      • Bachelor of Industrial Design2014 – 2016
        Raffles Design Institute
      • Micro- and Macroeconomics2012 – 2014
        South China University of Technology

      Languages

      • GermanNative
      • EnglishNative
      • HebrewFluent
      • ChineseConversational

      What employers say

      Translated from my German employment references. The original references are available on request.

      He stands out for his exceptional technical expertise, his strong entrepreneurial thinking and his outstanding ability to turn complex requirements into scalable, economically sustainable solutions on his own.

      Avi Ben-OrPresident, Austrian Small-Field Football Federation (ÖKFB)
      More from this reference
      • “The solutions he developed and implemented on his own have automated more than 80% of the routine administrative work that used to be done by hand.”
      • “The system landscape he created supports every level of the organisation and every user group, and allows a complex league operation with thousands of participants to run largely automatically.”
      • “In his board and leadership role he shows excellent organisational and decision-making skills. He sets clear priorities and ensures lasting project organisation through structured processes, targeted delegation and transparent communication.”
      • “He has always carried out every task he has taken on to our fullest satisfaction in every respect.”

      He impressed us with his excellent technical expertise, his strong analytical thinking and his entrepreneurial understanding.

      Michael DavidsonBridge Group
      More from this reference
      • “He stood out in particular for his high level of initiative, his structured and independent way of working, and his ability to master complex technical and organisational challenges.”
      • “His work made a significant contribution to modernising the hotel’s technical infrastructure and making its operation more efficient. The introduction of digital self-check-in processes in particular reduced the workload at the front desk.”
      • “He acted as the central interface between management, departments and external partners.”
      • “He always carried out all the tasks assigned to him to our fullest satisfaction.”

      Beyond work

      Sport, side projects and the things that got me into building software in the first place.

      • Grassroots football

        I volunteer on the board of the Austrian Small-Field Football Federation. Amateur sport only works because people give their evenings and weekends to it, and I like being one of them.

      • Building Goodfield

        On the side I am building Goodfield, a booking and management platform for padel, tennis and other sports clubs, so a club can run its own branded platform instead of renting space on a marketplace.

      • AI that watches the game

        I’m building AI match analysis for our league: shots, player movement and who’s who, from any recording or stream. My dream follow-up is an AI judge for trading card games that watches through an iPhone camera, checks plays and talks to the players.

      • A designer by training

        I studied Industrial Design before I wrote software for a living. It still shapes how I build: start from the person holding the thing, and leave out what they don’t need.

      • Mentoring

        At twinformatics I mentored seven junior developers and five apprentices in iOS development. Explaining my work to someone new is still the fastest way I know to find the weak parts of it.

      • Four languages

        German and English are both native for me, I speak Hebrew fluently and get by in Chinese, which I picked up while studying economics in Guangzhou.

      Building something that has to work on match day?

      I’m open to full-stack and AI engineering roles from December 2026, and always happy to talk about sports tech, vision AI or platforms that people rely on.

      LinkedIn Email