We're at the most important hiring moment in Kallidin's life. The first people we bring in will set the technical foundations, the culture, and the trajectory. We're not filling roles — we're picking the people who'll still be here when this is a platform used by hundreds of enterprises.
These are the areas we're hiring in — not a fixed headcount. Some will see multiple hires as we scale. If your experience spans more than one area, that's a strength. And if you don't tick every box on a list, apply anyway. The right mindset and the right experience rarely come pre-packaged exactly as described.
The most important technical hire we make. You'll design and build the Multi-Agent System at the heart of the ADO — the orchestration layer, the agent communication framework, the execution pipeline. Not a PoC. Not a prototype. A production system running on live enterprise data that has to be right at 3am when nobody's watching.
The problem is genuinely hard: how do you build a system of specialised agents that handles the full analytical workflow — business question to validated, executive-ready answer — reliably, at scale, on data environments you don't fully control? We have strong views on the answer. We need someone who can execute on them and isn't afraid to challenge them. In 18 months, the architecture decisions you make now will be running inside Fortune 2000 data teams.
Every enterprise AI failure traces back to the same root cause: the data underneath it was never trustworthy to begin with. The Unified Semantic Layer is how we fix that. It's the Rosetta Stone between fragmented enterprise data and the agents that query it — encoding business logic, metric definitions, and data lineage in a form that agents can interrogate reliably and a governance layer can audit adversarially.
You've done schema archaeology on real enterprise estates — the kind where "revenue" means six different things across eight systems. You know how to make it clean, versioned, tested, and trusted. This role is the foundation everything else is built on. Get it right and the ADO works. Get it wrong and it hallucinates. The quality bar is not negotiable.
The ADO's execution layer includes specialised predictive agents — demand forecasting, churn modelling, propensity scoring, anomaly detection. Your job is to build them and keep them honest. Production-grade pipelines, proper model governance, and a clear view of where ML belongs inside an agentic system and where it doesn't.
The interesting challenge here isn't the modelling — it's the integration. A prediction inside an agentic workflow has to be explainable to the adversarial governance layer, consumable by the narrative agent, and auditable by the client. Most ML engineers have never had to think inside those constraints. You will — and you'll shape how we solve it. If you've spent your career making models production-ready and you're frustrated that most deployments still treat you as an afterthought, this is built differently.
The consulting practice is the front door. It's where we establish trust, diagnose the structural problems, and design the foundations the ADO runs on. You'll be working directly with C-suite stakeholders — CDOs, CTOs, COOs — helping them understand what's actually blocking their AI programmes and what it takes to fix it.
This isn't a traditional consulting role. You won't be producing frameworks and slide decks and walking away. The diagnostic work you do becomes the blueprint for a technical build. That means you need genuine command of the data and AI landscape — enough to have a credible conversation with a Chief Data Officer about semantic layers and agent architecture, and enough commercial sense to frame it as a business problem rather than a technology pitch. At Kallidin, the consulting engagement is step one of a sequence that ends with a deployed AI system. You'll see the impact.
You're selling a category that doesn't have a name yet. Most enterprise AI projects fail — not because the technology doesn't work, but because the data infrastructure underneath it was never fit for purpose. Kallidin fixes that structural problem. Your job is to help C-suite executives see it clearly before they've named it themselves.
The model is diagnostic-first. The first engagement is a fixed-fee assessment — not a product pitch. You win by being right, not by being persistent. If you've built your career on relationships rather than cold calls, and you're frustrated selling AI solutions that don't fundamentally change anything, this is different. You'll be in the room with a CDAO talking architecture and a CFO talking payback period.
The founders have built and exited two data science businesses — including Aquila Insight, voted DataIQ's Best Place to Work in Data. We know what a great data culture looks like. We build it deliberately, not by accident.
If we aren't proud of the output, we don't ship it. Every deliverable that leaves Kallidin has been stress-tested before it reaches a client. The people who build here are proud of what they build.
We're an AI-native organisation. Flexible working patterns, flexible locations, and a comprehensive benefits package. We care about output, not where you sit.
We don't have a formal application process. Send us a note — who you are, what you've built, and why Kallidin. We'll take it from there.
If Kallidin sounds like the answer to a problem you're dealing with, start with a diagnostic. Two to four weeks to know exactly what's needed.
Start with a Diagnostic