About

I'm Joel Mercer. I build AI-native platforms from scratch and scale them up, always at the intersection of AI, data, and product. My work is about taking messy, conflicting signals and turning them into decision and narrative intelligence: clear, evidence-backed answers to high-stakes questions.

The unusual part of my profile is that I don't stop at strategy. I set product direction and AI architecture, and I stay hands-on in the build: agentic AI, LLM pipelines, React, TypeScript, Python, FastAPI, knowledge graphs, and secure deployment. The person who owns the roadmap is the one who can write the pipeline.

Logically.ai

I joined Logically as the first senior product hire and set up the product function from scratch. Twelve months later I was VP of Product, and over the next few years I led product, engineering, and data science as we scaled from an early platform to £8M ARR across government, defence, national security, and enterprise, from seed to Series A, contributing to $24M raised.

Along the way we shifted manual threat-detection workflows to agentic, AI-assisted intelligence operations and cut analyst triage time by 60%; ran a multi-tenant Google Cloud platform with Databricks-based data engineering and MLOps; and delivered cloud, on-premises, and air-gapped deployments that met government accreditation without fragmenting the core product. We helped turn narrative intelligence from a vague problem into a recognised market category, with a top-right placement in Gartner's Emerging Market Quadrant. The work was recognised outside the company too: number two on Fast Company's Most Innovative Companies in AI (2021), and a finalist for the UK's Most Innovative Cyber SME (DCMS & Infosecurity Europe). Most recently I designed and built PRISMα, a modular AI threat-detection and decision-intelligence platform.

Before that

My product work is rooted in behavioural research and design. I led UX and product engineering for multinational brands including Walmart, built data-visualisation and data-fusion tools for the NHS, Transport for London, and local government, and started out as a digital designer. That grounding, decisions based on observed behaviour rather than assumptions, still shapes how I build.

Now

I'm cofounder of VIXR, an LLM-driven volatility-intelligence platform for the S&P 500, and I take on a small number of fractional, advisory, and build engagements. I write about narrative intelligence, OSINT-grade tooling, and what it takes to ship AI into high-stakes, high-scrutiny environments.

What I work with

Product & strategy

AI product strategy, 0→1 development, B2B/B2G SaaS, go-to-market, category creation, fundraising, competitive analysis, Lean Startup

AI & decision intelligence

Agentic AI, LLM applications, RAG, NLP, knowledge graphs, entity resolution, model evaluation, explainable AI, human-in-the-loop workflows

Engineering

React, TypeScript, Python, FastAPI, rapid full-stack prototyping, API & platform design, behavioural UX

Platform & deployment

AWS, Google Cloud, Databricks, Spark, Delta Lake, vector search, MLOps, CI/CD, multi-tenant SaaS, observability. Cloud, on-prem, and air-gapped.