Eidetic

Robin Miklinski Independent AI product and systems engineer · London

I engineer intelligent systems that work in the real world.

I shape the architecture, data, integrations and controls around AI—turning promising prototypes and fragmented workflows into useful, dependable systems.

Explore the two routes → Start a conversation →
01 / Two routes

Two routes into one practice.

Building an AI product

For founders and product teams with a promising prototype that now needs production architecture, secure data boundaries, dependable behaviour and senior hands-on technical direction.

Applying intelligence to a business

For established businesses with valuable but fragmented workflows, disconnected systems or ageing software—and uncertainty about where AI would create genuine value.

02 / Selected proof

Selected proof.

Three records of turning ambiguity, complexity and risk into systems that hold up in use.

Reward Gateway Employment history

Eight years at Reward Gateway

Over eight years at Reward Gateway, I built the quality function from zero to fifteen engineers, supporting more than two hundred people in product and engineering. I designed the automation and measurement platform underneath it, coordinating roughly a thousand end-to-end tests across pull request, release and nightly tiers.

Later, I took hands-on responsibility for quality and release assurance on the company’s first customer-facing AI product: a multi-tenant retrieval assistant built on AWS Bedrock with OpenSearch across web, iOS and Android. I built the evaluation and adversarial security testing from nothing, covering answer quality, prompt injection, tenant isolation, personal data and retrieval scope, and gave the go/no-go panel evidence it could act on.

That work also included turning 451 production incidents into a £227,000 annual business case for preventable failure, and building a GenAI pull-request review agent that I presented at BrowserStack World Tour in London in 2024.

Architecture · Applied AI · Reliability · Technical leadership Employment history
YR2L, We Can Help You Limited 2025–2026

24,000 applications rescored in production

I took over a Dart, Flutter and Firebase product from its original solo developer, working from a codebase with almost no documentation. I integrated a third-party planning-data API, restored a failed council dataset, rebuilt the scoring system and re-scored more than 24,000 planning applications in production. I then stabilised the handover and onboarded the external development team that took the product forward.

Delivery Client engagement
Eidetic 2026

Original systems with explicit boundaries

I build and use agentic systems that keep human authority explicit: a private working environment that denies risky operations by default, a public daily workflow template for coding agents, and a draft-only email assistant that cannot send generated replies without human review. The same discipline shapes this site: ambitious systems, inspectable decisions and explicit release boundaries.

Agentic systems Inspectable and private systems

Open the complete evidence register →

03 / How I work

Lead, contribute and bring in specialists.

I take responsibility for the whole intelligent product. Depending on the work, that means:

  1. 01 Architecture, data and integrations

    Shape the system boundary, data flows and connections to the products and services around it.

  2. 02 Security, evaluation and reliability

    Define what dependable behaviour means, test the failure modes and make remaining uncertainty visible.

  3. 03 Hands-on technical direction

    Make the consequential engineering decisions and contribute directly where they matter most.

  4. 04 Leading client engineers while contributing directly

    Increase the capability of the team already closest to the product instead of creating a parallel delivery silo.

  5. 05 Bringing in specialists when the work requires them

    Add focused depth without losing one coherent architecture or line of accountability.

Output is abundant. Confidence is not.

04 / Background

Systems work, from traceability to intelligent products.

Robin Miklinski is an independent AI product and systems engineer who turns promising ideas, prototypes and fragmented business workflows into complete, dependable systems.

My background spans defence and intelligence data systems, retail at peak scale, collaboration software and global SaaS used by millions. I have worked from formal traceability and change control through product engineering, platform reliability, applied AI and technical leadership.

My work spans Python, JavaScript, TypeScript and C#, from product implementation to the evidence needed for sound decisions. I hold a BSc (Hons) in Internet Engineering from the University of Exeter.

Outside work, I DJ and produce electronic music as Eidetic. I have been a resident at EGG London and hosted Late Night Relics on AAJA Radio.

05 / Contact

Start a project conversation.