Founding Engineer (Infrastructure) in London

The problem we're obsessed with: The best analytical work in the world is locked inside human heads and PowerPoint slides. It doesn't compound. It doesn't scale. Every engagement starts from zero. A market map built this year gets filed away. The judgment a senior consultant develops over a decade; about what questions to ask, where the risks hide, how to structure a narrative; disappears when they move firms. Extraordinary talent. Workflows that haven't changed in twenty years. Riplo is a software company. We build the operating layer that makes expert analytical work repeatable, scalable, and compounding; starting with private equity diligence, the most rigorous, high-stakes analytical workflow in finance.

What you will do:

• Build the AI infrastructure that everything runs on. You are not joining a team with a finished architecture. You are one of the first engineers; which means you design and own the systems that power every engagement we run. The agent pipelines, the data infrastructure, the evals framework. The choices you make in the next twelve months will be the ones we live with for the next ten years.
• Go beyond RAG. We are not building a wrapper around an LLM. We are building multi-step agentic workflows with reliable, enterprise-grade inference; systems that can ingest messy, heterogeneous data and produce outputs that a PE partner would stake a deal on. You design the architecture that makes that possible.
• Own the full AI stack. Data ingestion, chunking strategies, retrieval, agent orchestration, output validation, evals; you own it end to end. You make the calls on what gets built, how it scales, and how we measure whether it works.
• Build evals that actually matter. In our domain, hallucinations aren't just annoying; they're deal-breaking. You build the evaluation infrastructure that gives us and our clients confidence in every output. You define what good looks like and you make it measurable.
• Translate the domain into systems. You understand that private equity diligence has specific structure; the questions that matter, the documents that carry signal, the outputs that drive decisions. You build AI infrastructure that reflects that structure, not generic pipelines.
• Everything else that matters. At this stage, the job changes week to week. What stays constant: you are in the room for every critical decision, and you co-own what follows.

The mindset:

• Reliability over novelty. You care about systems that work in production, not systems that impress in demos. You understand that in high-stakes professional services, a 95% accurate agent is not good enough; and you build accordingly.
• Systems thinker. You think in primitives and composition, not features. You identify the fundamental building blocks, design for scale from day one, and build infrastructure that compounds; not pipelines that break.
• Owner, not executor. You do not wait for specs. You see what needs to happen and you make it happen. If something is broken and it affects the mission, it is your problem to fix; even if it is not your job.
• AI-native by default. You already build, deploy, and scale end-to-end AI agents in production. You are a power user of Cursor or Claude Code, constantly exploring new tools, and genuinely excited about how AI changes what is possible; not just what you build.
• High bar, low ego. You hold yourself to a standard higher than what is asked. You seek feedback, close loops, and when someone has a better idea you say so.

Who you are:

You are a backend and AI infrastructure engineer with deep experience in Python, LLM frameworks, and distributed systems. You have shipped end-to-end agentic systems in production (not just prototypes) and you have strong opinions on how they should be built. You have worked with PydanticAI, LangGraph, or equivalent orchestration frameworks. You understand retrieval systems, embedding strategies, and the tradeoffs that matter at scale. You have some exposure to how consulting or professional services firms actually operate. You understand why the domain is hard, and why generic AI tooling doesn't solve it. You have clear evidence of sustained high performance, inside or outside of work. We do not care about pedigree for its own sake. We care about what you have actually built and how fast you learn.

Our stack: Python, TypeScript, PydanticAI/LangGraph, AWS, Terraform, PostgreSQL, Modal.

Why this job, why now:

Most engineers who are right for this role are good at their current job. On track. The path ahead is clear. This is not that path. This is the moment before the category exists; when the infrastructure decisions you make about how AI-native analytical work should be built will be the ones the industry copies in five years. You will have real ownership of what gets built. Direct access to a founding team from McKinsey, BCG, DeepMind, and Hg who have lived the problem and are rebuilding it from scratch. No middle management, no pointless meetings; just building something that matters. If you want to do the best technical work of your career and have it actually matter; this is the role.

What we offer:

• Competitive salary and meaningful founding-level equity from day one
• More ownership and architectural influence than engineers with far more senior titles at larger firms
• Direct daily access to founders who have built and backed category-defining companies
• The chance to define what AI-native analytical infrastructure looks like; from the ground up, not a ticket queue

If this is the problem you want to work on, we want to hear from you.