Skip to content
Redeemer Pace Download CV

Your next Software Architect. Your next CTO. Your next Tech Lead. Your next Product Engineer. Your next Architect who ships with AI agents. Your next Backend Engineer. Your next Frontend Engineer. Your next Blockchain Engineer. Your next Game Developer. Your next App Developer. Your next Software Engineer. Your next Product Owner. Your next CEO. Your next Growth and Marketing Lead. Your next Tech Consultant. Your next Game Consultant. Your next Business Consultant. Your next Marketing Consultant

Hello! I’m Redeemer Pace. Let's get to know each other, shall we?

Software architect and founder. Game engines, payments and ledgers, a blockchain, AI in production and platforms for millions.Open to architect, engineering leadership, product, blockchain and applied AI roles, forward deployed included, full time or B2B.

Maltese citizen, EU work rights, open to relocation, available now.

Download CVEmail me

Who I am?

Redeemer Pace, Architect. Builder. Founder.

Architect. Builder. Founder.

I build game engines, payment systems, a blockchain, platforms for millions and companies of my own.

I am a software architect with 20+ years writing software, since I was 7, and 15+ years in the industry, most of them leading. I have built 14 startups, and products at the companies I have worked for reach 200M+ active users. I design systems that are not tied to a vendor or a framework, and I stay hands on while I do it.

Today I lead architecture at Conrad for a 90+ person delivery organisation, on a storefront in 16 markets at up to 7M+ sessions a month. Before that I took Chiliz's fan app for 1.5M+ users in 167 countries from critically unstable to the core its seven squads ship on.

As a founder I grew a social network for anime and gaming fans to 10M+ registered users, and ran Gods of Zushin, an MMORPG on a C/C++ engine of my own, as a live business. I co-founded CoinOn, a Web3 platform I took from zero to web, mobile and its own chain as CTO. I also built a gamified social platform on payments, double entry ledgers and an ads engine with its own auction.

I have worked across e-commerce, payments, iGaming, game publishing and Web3, on the engineering side and on the business side. I have put machine learning into production since 2019, at KPMG and in my own products, and today I build with AI agents inside rules, context and checks I write, with a person reviewing every change. That mix is why I can be useful to a CTO in the morning and to a marketing team in the afternoon.

I am looking for an architect, head or VP of engineering, product engineering, blockchain or AI engineering role where the problems are hard and the standards are high.

years writing software, since I was 7
20+
years in the industry, from my first company in 2010
15+
startups built
14
active users on products at companies I worked for
200M+
users in 167 countries on one platform
1.5M+
sessions a month at peak on another
7M+
  • Maltese citizen with EU work rights
  • Open to relocation
  • Remote, hybrid or on site, based on location
  • Speaks Maltese, English and Italian
  • Available now

red.pace.dev@gmail.com | +356 79323059

CV
Full résumé (14 pages)
Me
GOZ
SN

Ask the Terminal

Prefer the command line? Everything on this page is one command away, and ask puts your question to my AI agent, which answers from this site. Start with help.
visitor@redgame.devPress / or ` from anywhere
Welcome to redgame.dev. Start with help: it lists every command, from contact to a few surprises.
Or ask me to do something: red create app, red fix my broken vibe coded app, red migrate app, red enable ai in my company, red hire as cto.
Or ask my agent about my work, in plain words: ask what has red built with machine learning?
Press / or ` anywhere on the page to come back to this prompt.

What can I offer?

What I take on, grouped by area. Each card is work I have done for real, and each one has a direct line to me.

Architecture & Platforms

Software Architecture & Technical Leadership

I take ownership of the architecture and stay hands on while doing it. I work across engineering, product, QA and stakeholders up to CTO level, and I design so that a vendor, a framework or a cloud can be changed without a rewrite.

  • Target architecture, trade offs and a migration path
  • C4 diagrams, decision records and a numbered decision log
  • Adapter contracts for every external vendor, with business logic kept free of framework code
  • Infrastructure behind adapters, with a written cut over plan that protects data
  • Technical direction across several squads, with mentoring and review
  • A clean handover guide at the end
Email me about thisLinkedIn

Payments & Ledger Architecture

I design and build the money side of a product so that it adds up, survives retries and passes review.

  • Payment gateway that is not tied to one provider
  • Double entry ledgers with reconciliation
  • Payouts, multi currency and tax per region
  • Holds and wallet adjustments above a threshold need a second admin's approval
  • Idempotency and signed webhooks end to end
  • KYC, AML and PCI scope designed in from the start
Email me about thisLinkedIn

Migration & Modernisation

I plan and lead migrations off legacy systems, from the first audit to the last cut over. I have moved a storefront in 16 markets off AEM to a headless CMS, an unstable mobile app onto a new core, Backbone frontends to Redux and RxJS, brownfield client systems in finance, iGaming and pharmatech, and a Web3 platform from Polkadot parachains to a chain of its own.

  • An audit of what moves: components, content, data, integrations, redirects and translations, all counted
  • The target picked on evidence: a scored evaluation of platforms, frameworks or chains
  • An adapter so old and new run side by side behind a feature flag
  • A working proof of concept before commitment
  • Rollout per market or per slice, with monitoring and rollback
  • A business case that shows what the move saves, licence and running cost included
Email me about thisLinkedIn

Architecture & Security Review

I review a codebase or a set of pull requests and return findings against the real code with a fix plan.

  • Findings shown on the exact lines, with numbered fix steps
  • Authorisation, billing integrity and data exposure checks
  • Supply chain and dependency review
  • A definition of done the re-review is checked against
Email me about thisLinkedIn

DevOps & Release Engineering

I set up the pipelines, environments and release paths that let a team ship often without breaking production.

  • CI/CD with audit logs, permission controls and rollback
  • Kubernetes, canary rollouts and edge caching, or simpler hosts behind deploy adapters
  • Environments, migrations and feature flags that keep staging honest
  • Disaster recovery and business continuity plans
Email me about thisLinkedIn

My History

Employment runs on the main line. The companies I founded run beside it on their own branch, because I have always built alongside the day job.
Industries I know

Experience

  1. Nov 2025 to Present

    Lead Software Architect / Enterprise Architect, Conrad Electronic Group

    Up to 7M+ sessions a month, working with 90+ people up to CTO level

    Conrad Electronic is a European electronics retailer operating across 16 markets, DACH-led.

    • Designed and built the CMS Adapter API, an anti-corruption layer with handler registry, BFF pattern, versioned response contract, HMAC-signed authentication, and versioned preview tokens
    • Authored 11 numbered Architecture Decision Records covering adapter architecture, versioning, cache invalidation, BFF dispatcher, feature-flag integration, authentication, media handling, editor plugin, email templating, and trunk-based Strapi config
    • Kept the CMS adapter framework free by construction: 69 TypeScript files with zero Vue or Nuxt imports, one typed contract, Strapi, AEM and backend for frontend providers chosen by feature flag, and a block registry that uses TypeScript exhaustiveness so no block type ships unregistered
  2. Oct 2023 to Nov 2025

    Senior Full Stack Engineer, HyperPlay Labs

    HyperPlay is the world's first Web3 game launcher, pioneering how decentralized games are discovered, played, and integrated with crypto ecosystems.

    • Built scalable architectures for both the web platform and Electron based desktop app, ensuring maintainability and high performance.
    • Introduced and architected the mobile app using React Native, enabling a consistent cross platform user experience. English
    • Implemented and optimized scalable features across the main website, game store, developer portal, and launcher client.
  3. Jan 2023 to Aug 2023

    Senior Frontend Engineer, reNFT Labs

    An interesting company that offers NFT rentals, integrating with various industries like gaming.

    • Developed NFT Marketplace V2 in close collaboration with the Tech Lead, ensuring a scalable, modular architecture.
    • Architected and implemented performance optimized solutions, achieving high Lighthouse scores and following best practices for frontend performance.
    • Built reusable UI components using Radix UI, collaborating effectively with the Design team to maintain consistent design language.
  4. Nov 2019 to Nov 2022

    Mobile Core Engineer / Architect, Chiliz

    1.5M+ users in 167 countries, 7 product squads and 30+ engineers

    A blockchain and sports company building a gamified fan engagement platform with real-time rewards, NFT integrations, and immersive mobile and web experiences.

    • Joined as a Mobile and Frontend Engineer, quickly expanding into Tech Lead and Architecture roles due to early stage startup needs and team dynamics.
    • Built the V2 Core Architecture ("Epics Rx") from scratch, implementing routing, Sentry integration, app lifecycle handling (e.g., back press), and RxJS based state orchestration.
    • Architected the new V2 App foundation, introducing structure, modularization, and maintainability across mobile and web platforms in close collaboration with Tech Leads.
  5. Feb 2019 to Sep 2019

    Software Engineer / Technical Lead for Client Projects, KPMG

    Being hired by KPMG is one big achievement on its own, as it is one of the big 4 firm in auditing.

    • Contributed to and led architecture decisions across multiple brownfield migrations and greenfield projects for KPMG clients across finance, iGaming, Web3, game publishing, and enterprise software.
    • Developed and delivered full stack features across frontend, backend, and core software, taking ownership from requirements and architecture through implementation, testing, deployment, and client delivery.
    • Worked with multiple industries daily, often adapting rapidly to sector specific constraints and requirements.
  6. Jul 2017 to Jan 2019

    Frontend Game Engineer, AuthenticGaming

    One of the companies that given me a lot of experiences throughout the years. This company is an iGaming Company, which was my first experience in iGaming.

    • Went beyond the role of a Frontend Engineer, actively contributing to real time backend logic using Node.js and WebSockets to support live game data and streaming interactions.
    • Developed internal operator tools that streamlined and automated the onboarding and integration of new casino operators, reducing manual effort and scaling client support.
    • Contributed to the design and development of a new architecture, improving code maintainability and scalability across the frontend stack.

Education

  1. Sep 2017 to Present

    Online Courses

    These courses have helped me enhance further knowledge for both practical and theoretical

  2. Sep 2015 to Jun 2017

    Bachelor of Science (Honours) in Multimedia Software Development, MCAST

  3. Sep 2013 to Jun 2015

    Extended Diploma Computer Software Engineering, MCAST

  4. Sep 2011 to Jun 2013

    Diploma Computer Software Engineering, MCAST

  5. Sep 2010 to Jun 2011

    Foundation in Computing, MCAST

How I use AI day to day

I built machine learning models before LLMs and used AI coding tools before ChatGPT, and I have built integrations and internal tooling on the GPT, Claude and Gemini APIs for content, classification, data extraction and agent workflows, across my own products, client work and Conrad. Today AI is part of almost every task I work on, from writing code to preparing a stakeholder update. Here is what that looks like in numbers, the rules I work by, and how I got here.
At least 33,000 prompts to my main coding agent on one machine alone, February to October 2026. ChatGPT, Gemini, other tools, other machines and earlier years are not counted, so the real total is several times higher.

AI is in the loop on all of it

Where 33,000+ prompts went, on one machine, February to October 2026.

  1. Feature implementation and coding7,900+
  2. Code review, PRs and git3,400+
  3. Debugging and fixing3,050+
  4. Docs, stakeholder updates and decks2,800+
  5. Deploys, infrastructure and migrations2,200+
  6. Data, analytics and research1,750+
  7. Testing and verification1,450+
  8. Architecture and planning1,350+
  9. Security and compliance1,050+
  • Counts come from sorting prompt text by keyword, so they are approximate floors. A prompt can count in more than one row, so the rows do not add up to the total.
  • About half of all prompts are short steering turns like "continue" or "recheck", which sit outside these rows. The count is one coding agent on one machine since February 2026. ChatGPT (much of my architecture research and stakeholder writing), Gemini, other tools, other machines and the years before are not in it, so the real total is several times higher.
  • The mix moves with the work: architecture and consulting lean towards docs and stakeholder updates, and a verification heavy branch leans towards testing on the deployed build.

Tokens as a budget

I treat tokens as a budget. About 96% of the input in my agent sessions is served from cache, small models do the searching while large models make the decisions, and agent runs are capped and resumed instead of relaunched.

of input tokens served from the prompt cache
96%
output tokens
53M+
models across three tiers
8

Output by model tier, roughly: the right model per task, mostly the largest

  • Largest tier, about 85%Design, review and long implementation work
  • Middle tier, about 8%
  • Newer top model, about 5%
  • Smallest tier, about 2%Search, gathering and audits

How the budget holds

  • Large models for design and review judgement, small models at low effort for search, gathering and audits
  • Multi agent runs capped at about five agents and run one at a time, after one 59 agent run used a whole session quota
  • A stopped agent is resumed, not relaunched, so it does not read everything again
  • No builds or type checks after every edit and none inside subagents: verification runs once, before a push
  • Trimmed transcripts, digests and search, so old context costs nothing until it is needed
  • Skills and commands, so the agent does not spend tokens rediscovering the codebase every session
  • Measured on one machine, February to 1 August 2026. Cached input is billed at a fraction of fresh input, which makes the cache the biggest cost lever in long agent sessions.
  • At Conrad I work on AI enablement under budget constraints across teams and departments.

Subjects I use it on

The same prompts sorted by subject instead of by kind of work. Subjects overlap, so these do not add up to anything.

Over a thousand prompts each

  • Database and migrations
  • Testing, unit to end to end
  • Ads and feed ranking
  • UI and design systems
  • CMS and content modelling
  • Stakeholder docs and decks
  • Authentication and identity
  • CI/CD and hosting

Hundreds of prompts each

  • Data analysis and audits
  • Architecture and design docs
  • KYC, AML, tax and legal compliance
  • Ledgers, wallets and treasury
  • Payments and subscriptions
  • Performance and optimisation
  • Analytics and tracking
  • Caching and CDN
  • Blockchain
  • Cost and vendor evaluation
  • Translations across locales
  • Observability
  • Mobile and offline
  • Security hardening
  • Real time streams
  • Developer platform and SDK design
  • SEO
  • Performance profiling and optimisation, especially for canvas and other compute heavy apps, has no row in the task split above: it is usually phrased as a bug, a test or a refactor, so a keyword split cannot isolate it.
  • Canvas and game work barely shows here because most of it happened on my other PC, which is not counted.

How I work with agents

Every change goes through the same four stages, and each stage has rules the agent cannot skip.

  1. 1Context

    • The contract matters more than the promptEach repo has a context file with hard rules, reference docs the agent reads first, a memory of one fact per file and a handoff note per session.
    • Confidentiality firstI check a tool's confidentiality model before it touches company code. Every tool an agent gets is scoped to read or write, secrets are checked by name only, and no sensitive data goes to the model.
    • Context that outlives the sessionI wrote tooling that trims multi gigabyte transcripts without deleting them, builds a digest and topic index per archived session, searches across all of them, and serves that history through a read only MCP server with its own evals.

  2. 2Agent

    • Production is a deliberate stepStaging and production are separate connections I choose explicitly. Subagents never write to production, and migrations are applied only on my go and proven afterwards.
    • Synthetic data at real volumeMoney and customer paths run on synthetic data at production volumes, so scale problems show up before release.
    • Agents with limitsSubagents research and edit but never build or commit; the main session builds once. Parallel runs are capped at about five agents, after one large run used a whole session quota.

  3. 3Checks

    • Dependencies are screenedEvery install passes a supply chain blocklist, and a match fails the install.
    • Proof before acceptanceWork is accepted on evidence from the deployed build, such as Playwright runs and API responses, with test data created through the real API instead of seeded.

  4. 4Human review

    • A person reviews every changeNothing merges without a human review, and features nobody asked for come back out.
    • Decisions are written downEvery open choice comes to me as a question with a recommendation and trade offs, and my answer is recorded as a numbered decision. One branch holds over 800 of them.

Caught in review

  • A pattern violation in the generatorNine domain services used a shared base class directly instead of a subclass of their own. Each was wrapped in its own class, and the code generator template that kept emitting the violation was fixed too.
  • Live secrets marked as unusedRepo instructions called two secrets dead and said to delete them. They were still live on the demo and production deploys. "No longer used" now means on every deployed branch.
  • One listener, registered twiceA review page showed that a new pull request had absorbed an older open one. Merging both would have registered the same listener twice.

The same setup serves engineers with reviews, migrations, tests and deploys, and it serves product and analysis stakeholders too.

How I got here

  1. Up to 2019

    Machine learning before LLMs

    From personal TensorFlow experiments to shipped models for the clients KPMG placed me with as a consultant: suggestions and recommendations, ads, and social media feed ranking.

  2. Before ChatGPT

    Tabnine

    Code completion with Tabnine gave me an early, realistic sense of what AI assistance could and could not do.

  3. Early 2023

    ChatGPT at reNFT

    I started using ChatGPT seriously for frontend and backend work, from authentication and API design to serverless architecture. I wrote a small CLI that built a short context file for a project (architecture, directory layout, key files), choosing what went into it with company safety and privacy in mind. The rest was manual: paste the context, adjust, create the files by hand.

  4. Later in 2023

    Cursor, Claude and Gemini

    I picked each one up as it appeared and have kept current since.

  5. Oct 2023 to Nov 2025

    HyperPlay

    AI across frontend, backend, Electron and mobile work.

  6. Feb 2025

    Project instruction sets

    I started writing detailed instructions for AI assisted development: architecture constraints, security requirements and conventions, so the model works inside the project instead of guessing at it.

  7. Nov 2025 to present

    Conrad Electronic Group

    I drive AI enablement for the CMS migration, working with 90+ people across engineering, QA, product, content, SRE and up to CTO level. That covers Gemini Enterprise and Gemini CLI as the company tooling, a presentation to the team on how MCP and skills help with Gemini and Claude models, an evaluation of Strapi's MCP server and of the AEM MCP options (I recommended against the AEM ones), a supply chain check across the storefront repositories after the axios compromise, written up as an advisory for the team, and a human review before anything AI generated reaches Confluence or a pull request.

  8. 2026

    My own platform

    Claude Code is my main engineering tool on a large TypeScript monorepo for payments, ledgers, an ads engine and real time feeds. Around it sit skills adapted from an open source collection and extended with my own, such as an engineering discipline skill organised as phases and gates, a design review subagent, and private published pages that keep the working record: decision registers, review pages teammates use as their fix list, and live trackers.

  9. Ongoing

    Gods of Zushin, my MMORPG

    I use AI to build features, write up in game events and debug support cases. For support I copy only the non sensitive data for that case into a sandbox of my own and let the agent work there, never on live data. Before off the shelf MCP servers existed I built my own for logging, audit and SQL access; now I use and configure the published ones. In September 2026 I also ran an owner authorised, read only security assessment of its public surface and wrote the hardening runbook.

On Chain

Web3 from fan tokens at scale to a chain of my own. Each block below is somewhere it shipped, confirmed as you read it.

Web3 infrastructure, exchanges and asset management: from token trading at scale to a chain of my own, lending, and wallets.

Where it shippedChilizCoinOnreNFTHyperPlay

The stack on chain

Token standards

  • ERC-20Fungible tokens and stablecoins
  • ERC-721NFTs, one of a kind
  • ERC-1155Many token types in one contract

Smart contracts

  • Solidity
  • Rust on Substrate
  • Security audits
  • Attack path review

Chains

  • Ethereum and EVM chains
  • Polygon
  • Solana
  • Polkadot and parachains
  • Substrate

Wallets and chain data

  • MetaMask
  • WalletConnect
  • Wagmi
  • Ethers.js
  • Web3.js
  • The Graph
  • Alchemy
  • Infura

Send a transaction through the stack

Pick a request and watch it travel from the wallet to the chain and back, with security checks on the way.

A simulation of the architecture: nothing is signed and nothing touches a real chain.

  1. WalletMetaMask, WalletConnect
  2. dAppWagmi, Ethers.js
  3. Security checksSimulation, Allowance rules
  4. RPC providerAlchemy, Infura
  5. MempoolNonce, Gas
  6. BlockConfirmations
  7. IndexerThe Graph
  8. UIReceipt
  1. Block #0001Pending

    Token exchange and fan tokens

    An exchange app where each club's token trades against a base token, with on chain buying and reward claims, for 1.5M+ users in 167 countries.

    • Chiliz
  2. Block #0002Pending

    A chain of my own, with bridges

    A chain in Rust on Substrate after Polkadot parachains, with bridges and indexers across Solana, Polkadot and EVM chains.

    • CoinOn
  3. Block #0003Pending

    Stablecoins

    Stablecoin work as co-founder and CTO, for clients, and on my own platform.

    • CoinOn
    • Client work
    • Own platform
  4. Block #0004Pending

    DeFi and trading

    DEX and AMM, lending and staking, and a front end for configuring and deploying grid trading strategies on an on chain order book, with MetaMask and WalletConnect.

  5. Block #0005Pending

    NFT lending and marketplaces

    The NFT lending protocol and its marketplace V2, built with Wagmi and Ethers.

    • reNFT
  6. Block #0006Pending

    Web3 game distribution

    A Web3 game launcher with wallet connections across EVM chains.

    • HyperPlay
  7. Block #0007Pending

    Wallets and account flows

    A wallet platform on Polygon: wallet sign in, NFT and token balances, ERC-721 and ERC-1155 transfers by username.

    • Own project
  8. Block #0008Pending

    Indexing and chain data

    Chain data through The Graph subgraphs and Alchemy.

    • reNFT
    • CoinOn
    • HyperPlay
  9. Block #0009Pending

    Contract security review

    Solidity reviews and smart contract security audits, looking for attack vectors before they ship.

    • HyperPlay
    • CoinOn
  10. Block #0010Pending

    Token economy design

    An off chain credit and coin economy: an append only ledger, a bonding curve, coin launch and trading.

    • Own platform

More Skills by Area

Areas I build in beyond the skills in the forge. Each comes from several roles, ventures and client projects, my own platform being one of them, so they are listed rather than rated.

Payments & Fintech

  • Stripe, including Connect payouts
  • PayPal for purchases and donations
  • Subscriptions
  • Payment provider abstraction and routing
  • Double entry ledgers
  • Idempotency
  • Signed and verified webhooks
  • Reconciliation
  • Multi currency and FX
  • Tax
  • Banking rails: bank transfers next to card payments
  • On and off ramps between fiat and crypto
  • Payments paused automatically on fraud signals
  • Payment holds with dual approval

The areas at work together

Three systems from the multi market commerce platform I work on now: how it meets outside systems, how it ships market by market, and how it is watched with the shopper's consent. Each is a glance, not the full design.

What I Have Built, More Than Once

The same kinds of system, built at several companies. Each tile names where; below them, worked examples in detail.

  • Payments and ledgers

    Gateways, subscriptions, double entry ledgers, and regulated deposit and withdrawal flows.

    • Own platform
    • KPMG clients
    • Gods of Zushin

  • Banking rails and ramps

    Card and bank payments side by side, fiat to crypto and back, and payments paused automatically on fraud signals.

    • Own platform
    • CoinOn

  • Exchanges and trading

    Trading and swaps from one token to another, listing after KYC review and automatic listing on volume.

    • Chiliz
    • CoinOn

  • Tokens and launchpads

    Fan tokens at scale, creator tokens traded against a pegged stablecoin, and NFTs.

    • Chiliz
    • CoinOn
    • reNFT

  • Chains, wallets and indexers

    Multi chain apps with wallets, indexers and bridges, and a chain of our own on Substrate.

    • CoinOn
    • HyperPlay
    • reNFT
    • Chiliz

  • Virtual economies and loyalty

    Currencies earned by taking part, loyalty points, rewards and in game shops.

    • AMW
    • Gods of Zushin
    • Punti
    • Chiliz

Kinds of architecture I have built

  • Full financial systems, fiat and on chain
  • Double entry ledgers
  • Sagas across ledgers
  • Microservices
  • A modular core behind adapters
  • Event driven, with an outbox
  • Backend for frontend and anti corruption layers
  • Micro frontends
  • SSR, SSG and edge caching
  • Offline first, with sync
  • Real time over WebSockets and SSE
  • Client and server games for low latency
  • Multi chain, with bridges and indexers
  • A chain of our own on Substrate
  • Monorepos with a shared core
  • Machine learning pipelines

Worked examples

A full financial system end to end, multi market commerce, my own platform's core and one payment through it, a ranking pipeline and an ads engine. Each is a glance, not the full design.

Code Review and Open Source

The work behind the work: reading other people's code closely, in teams and in the open.

From one GitHub account alone, so it is a floor: most of my review work at Conrad, Chiliz, KPMG and Authentic Gaming lived in company repositories this cannot see.

Pull requests merged

One square per pull request

Pull requests reviewed

One square per pull request

Activity on GitHub

The last three years on my current account, darker for busier days. The two decades before live on earlier accounts and in company repositories.

  1. 2024
  2. 2025
  3. 2026

Each year's longest streak, outlined

Achievements unlocked

  • 37 day streak
  • 39 weeks in a row
  • 559 active days
  • 20 perfect weeks, all seven days
  • Busiest month: March 2026

fetchff

An open source HTTP client library. It is not mine: I review its pull requests, 35 so far, and opened one of my own.

I review a library my own platform depends on instead of forking it.
  • Its security hardening release, with full test coverage
  • React Native support and high throughput performance
  • SWR support and a React hook
  • Request deduplication and better polling
  • Retry with jitter, limited to idempotent methods
  • Cache revalidation with ETags, and request aborting
  • Endpoints typed from OpenAPI schemas
fetchff on GitHub

Issues I opened

43 issues in other teams' public repositories, from security findings to component specs and bug reports.

  • 13 security findings from a code scanning and dependency audit of a desktop game launcher
  • Prototype pollution, ReDoS, command line injection and CSRF among them
  • Component specs and fixes for a design system's UI library
  • A bug report on Prisma running on Cloudflare Workers
My GitHub

Others

Open source projects and the company codebases I have worked in.

  • 91 pull requests reviewed on my own platform's repository

Live Table

iGaming across live casino, slots and bet tables, to regulated platforms for clients. The cards are dealt as you arrive.

iGaming from the inside: live casino, slots and bet tables, player accounts and wallets, bonuses and tournaments, sportsbook and live odds, game aggregation, and regulated platforms for clients.

Live

  • Live casino UI

    Game UI for desktop and mobile live casino tables, and the mobile app.

    Authentic Gaming

  • Real time

    Node.js and WebSockets for live game data and streaming, with client side state kept in sync.

    Authentic Gaming

  • Operator integration

    Internal tools that automated onboarding and integration of new casino operators.

    Authentic Gaming

  • Tables on canvas

    A PixiJS bet table proof of concept, presented to the engineering team and the CTO.

    Authentic Gaming

  • Canvas performance

    Profiling tools that found the bottlenecks in canvas rendering.

    Authentic Gaming

  • Player accounts and wallets

    Accounts, balances and one wallet every bet and win moves through, with deposits and withdrawals under licence.

    KPMG, Authentic Gaming, My crypto casino

  • Bonuses and tournaments

    Bonus rules with wagering tracked across games, and tournaments with leaderboards and prizes.

    KPMG, My crypto casino

  • Sportsbook and live odds

    Live odds fed in and moving with the match, prices checked as a bet is accepted, and bets settled from results.

    KPMG, My crypto casino

  • Game aggregation

    Games from many studios behind one integration, so an operator adds a studio without new work.

    Authentic Gaming, KPMG, My crypto casino

  • Crypto deposits and withdrawals

    A crypto cashier on several chains: an address per player, confirmations that credit the balance, and checked withdrawals.

    My crypto casino

  • Regulated platforms

    iGaming clients needing compliance, scale and real time performance, including regulated deposit and withdrawal flows under Malta licensing.

    KPMG

Dealt atAuthentic GamingKPMG
As a software engineer, Redeemer would be a true asset to that position and it comes with my heartfelt recommendation.Head of Frontend, Authentic Gaming

Characters I Play

The roles I play, as characters in my own party, each with the job titles it covers. Level is the years in that role, measured from real dates. Stats are my own scores.

CTO

Chief Technology Officer, Head of Engineering, VP of Engineering11 years in the role, since 2015
  • Architecture95
  • Platform Engineering85

Abilities

  • Technology roadmap
  • Cybersecurity
  • Infrastructure
  • Executive team

Hands on as CTO: at CoinOn I set the roadmap and built the stack from 0 to 1 with a team of 10+, writing code, not only plans.

1 of 10

Skills

No self ratings. Each skill's rarity is earned from how long I have used it in real roles, projects and study, measured from their dates, and every card says where.
  • Legendary 8+ yrs
  • Epic 5+ yrs
  • Rare 2+ yrs
  • Common Hands on

Programming Languages

The languages I write, from C/C++ game engines and PHP frameworks of my own to TypeScript across a whole platform.

  • HTML / CSSLegendary16 yrs, Conrad Electronic Group, HyperPlay Labs +13
  • JavaScriptLegendary16 yrs, Conrad Electronic Group, HyperPlay Labs +15
  • PHPLegendary16 yrs, Conrad Electronic Group, Gods of Zushin +4
  • PythonLegendary16 yrs, Gods of Zushin, KPMG +1
  • C#, .NETLegendary15 yrs, Gods of Zushin, KPMG +4
  • SQLLegendary14 yrs, HyperPlay Labs, Gamified Social Network Platform +3
  • BashLegendary11 yrs, Gods of Zushin
  • C/C++Legendary11 yrs, HyperPlay Labs, Gods of Zushin +4
  • JavaLegendary11 yrs, Chiliz, MCAST +1
  • LuaLegendary11 yrs, Gods of Zushin
  • TypeScriptLegendary9 yrs, Conrad Electronic Group, HyperPlay Labs +20
  • LESS / SASSEpic6 yrs, Conrad Electronic Group, HyperPlay Labs +2
  • ShaderLabEpic6 yrs, Engine Prototypes, Three Unity Games
  • PL/pgSQLEpic5 yrs, Gamified Social Network Platform, Freight Delay Notifications
  • SolidityEpic5 yrs, HyperPlay Labs, reNFT Labs +2
  • DartRare3 yrs, Chiliz, CoinOn
  • KotlinRare3 yrs, Chiliz, CoinOn +1
  • SwiftRare3 yrs, Chiliz, CoinOn
  • RustRare2 yrs, HyperPlay Labs, CoinOn +1
  • GolangCommon2 mos, CoinOn

Talents

The skills that do not fit in a stack: how I work with people.

Passive talents

  • Team work
  • Face to Face Communication
  • Remote Communication
  • Mentoring
  • Managing
  • Company Culture Follower
  • Contribute Ideas
  • Executive Stakeholder Communication
  • Written Documentation & Architecture Communication
  • Cross-functional Collaboration
  • Strategic Decision-making
  • Async / Documentation-first Working
  • Leading squads and engineering teams
  • Hiring and building teams
  • Architecture governance
  • Roadmaps and prioritisation
  • Estimation and scoping
  • Onboarding engineers
  • Technical and vendor evaluations
  • Delegation and ownership
  • Running workshops and discovery
  • Leading through incidents

Languages

  • MalteseNative
  • EnglishFluent
  • ItalianFluent

Boss Fights

Real problems from across my roles, client work and my own platform. Each boss loses health as you read how it was beaten, and drops the rule I kept.
Show case studies for
BossArchitecture & cost

A migration that pays for itself

Built the business case and the architecture to cut an enterprise platform's yearly licence and running cost by about 80%, infrastructure included.

  • A target architecture with the old and new platforms side by side behind one adapter, switched per market by a feature flag
  • A working proof of concept before commitment, so the case rests on running software instead of slides
  • Licence, hosting and infrastructure counted for both the old and the new stack
Loot: A business case that rests on a working proof of concept
  • Business case
  • Migration
  • Adapters
1 of 24

Human vs Agent

PvP, honestly: every round here happened. Most go my way. Some go to the agent, and each of those left me a rule I still follow.
AgentMeRounds won 4 of 7
  1. Round 1
    AgentA payer can change their country to unlock instant statutory refunds.Overruled
    MeThe country locks once set. The real gap is that it is never checked against the payment.

    Shipped: A jurisdiction snapshot on every charge.

  2. Round 2
    AgentThe configuration says otherwise: a production database connection exists without a read only flag. The absolute claim would not survive one question.
    MeWrite that agents never get write access to production and never see production data.Overruled

    Shipped: The accurate version: staging and production as separate, chosen connections, and subagents that never write to production.

    Rule kept: A claim about how I work is checked against the actual configuration before it is published.

  3. Round 3
    AgentThese two auth secrets match no code. Delete them from both hosting projects.Overruled
    MeDemo and production still run the old provider. Deleting them takes production down.

    Shipped: Live secrets kept, and "no longer used" now means on every deployed branch.

  4. Round 4
    AgentTwo kinds of query: one follows a subscription's own link to its plan, where a filter can only lose data; catalogue browsing keeps it.
    MeKeep the role filter on plan lookups; without it, how do we filter plans by role?Overruled

    Shipped: The split, after a fee bug charged some accounts the default rate instead of their plan's. I added the case it had missed.

    Rule kept: When a link already says which row is right, do not filter it again.

  5. Round 5
    AgentNine domain services can use the shared base class directly.Overruled
    MeEach domain gets a subclass of its own, and the generator stops emitting the shortcut.

    Shipped: Nine services wrapped and the generator template fixed.

  6. Round 6
    AgentThe documents mix a two year total with one year, and the public record for those years shows fewer users.
    MeA cost saving and a user figure, written as I remembered them.Overruled

    Shipped: About 80%, worded as the case and the architecture, and 1.5M+ users in 167 countries.

    Rule kept: Floors only, the measure named, nothing rounded up.

  7. Round 7
    AgentPost the review with the page's finding numbers.Overruled
    MeComment on the exact lines instead.

    Shipped: 43 line comments, and three fix samples calling helpers that do not exist caught before posting.

Rules of engagement

What I put in place before a team works with agents, so the speed comes with guardrails.

  • Security

    Tools scoped per environment, with staging and production on separate connections. Subagents never write to production, secrets are handled by name only, and a supply chain blocklist runs in CI.

  • Company policy

    What employees and contractors may use AI for, written down. At Conrad I enforced the rule that nothing AI generated reaches the wiki or a pull request without human review.

  • Data privacy

    A tool's confidentiality model is checked before company code goes near it. No sensitive data to a model, and synthetic data for money paths.

  • The right skills and MCP servers

    Project skills and MCP servers installed per job, each evaluated for access control before the team relies on it.

  • A process that scales

    One flow from idea to code: plan, design doc, tickets, test first, review. Decisions go in registers the agents search before they ask.

  • Fits the frameworks you run

    Built to plug into what a company already uses: Agile sprints, the issue tracker and the wiki, through to HR and accountancy processes.

Projects & Achievements

Every project is a region on the map, laid out in the order I explored them. Pick one to open its map: what I built there, the loot, and where to see it.
Show regions

Engine Room

Under the floor of the game world: an MMO from engine to live operations, a location based mobile game, rendering heavy animation at frame rate, a game server before and after a bot flood, and a Web3 game launcher.

Recommendations

Lines from my reference letters from Authentic Gaming and reNFT, attributed by role. The full letters from Authentic Gaming, KPMG and reNFT are available on request.
The breadth of his knowledge about software, systems and architecture is excellent and has shown to be valuable beyond the scope of front end work.
Lead Front End DeveloperreNFT, August 2023
Redeemer is a solver of problems at heart and an absolute pleasure to work with.
Lead Front End DeveloperreNFT, August 2023
He had a crucial role in the development of the game UI for the desktop and mobile application enabling AG to deliver the most innovative User Experience in the industry.
CTOAuthentic Gaming, January 2019
Redeemer demonstrated to understand the business requirements fully and to implement the solution as per the needs of the technical team.
CTOAuthentic Gaming, January 2019
Redeemer's efficiency and thrive to grab multiple projects and give them the perfect 'bang', is outstanding.
Head of FrontendAuthentic Gaming, January 2019
As a software engineer, Redeemer would be a true asset to that position and it comes with my heartfelt recommendation.
Head of FrontendAuthentic Gaming, January 2019

Bug Raid

A break from reading. Squash the bugs before they reach production. Your best score stays in this browser.
Score
0
Lives
3
Wave
1
Best
0

Bugs are heading for production.

Click or tap a bug to squash it, or use the arrow keys to aim and Space to squash. Regressions take two hits and flaky bugs jump.

Run complete

Wow! You made it to the end.

It launches by itself when you get here. To launch it yourself, hold the button, or press it once. Space and Enter work too.

Your run, lit up on the way out

Zones crossed
0/6
Bosses defeated
0/24
Duels played
0/7
Character
None picked
Bug Raid best
Not played
Time on this run
0:00
Rank: Explorer0 of 5 objectives

Next mission

Work that matters for people, and the systems that take us further: AI that is safe and useful, and the software that gets us off this planet.

Final quest: start a conversation

How did you find the journey? Write to me anytime in any timezone.

Email meLinkedInDownload my CVFull résumé (14 pages)

Not hiring right now? If you know a team that is, I would be grateful if you passed my name on.Share the site on LinkedInWorked with me? Recommend me on LinkedIn

Like the site?

Watch it being built: every commit since the first, replayed as a city that grows. Each building is a part of the codebase, as tall as its lines of code.
The site started in 2022 and sat mostly still until 2026. The rebuild came after, with agents working under the rules, checks and reviews described in How I use AI.
Accessibility
Follow Me
Pick a characterXPBosses defeated: 0/24