Founding Engineer

Who's Mubit?

AI agents are being deployed into production faster than the infrastructure can support them. The result: engineers spending 5+ hours a week manually compensating for agents that have no memory of what they've done, what failed, or why.

Mubit is the operational memory and state management layer for production AI agents. We give agents persistent execution history, shared state, and ACL-gated memory — so they get more reliable over time without constant human intervention.

We're well funded and building the founding team. This is your chance to get in at the beginning of something real.

You’ll be the first engineering hire and co‑own the architecture of Mubit's core infrastructure.

Mubit sits in the hot path of production AI agents — sub‑80ms retrieval, ACL‑gated shared state, pub/sub event triggers, execution history across agent fleets. The systems you build are why agents get more reliable over time instead of requiring constant human intervention.

This isn’t a spec‑execution role. You’ll identify the problems, own the trade‑offs, and ship the solutions.

Our office is in Canary Wharf (WeWork - 30 Churchill Place).

• Retrieval latency is consistently under 80ms at real customer scale
• The SDK is something engineers recommend to each other — not something they tolerate
• You’ve made architecture calls we’re still building on at 10x current load
• Early customers have integrated Mubit into production agent stacks without hand‑holding

• Get deep into the core storage and retrieval architecture — understand every trade‑off already made and form opinions on what to change
• Work directly with early customers integrating Mubit into their agent stack (Lang Graph, CrewAI, Auto Gen, etc.)Profile latency, find the bottlenecks, fix them
• Contribute to the SDK engineers use to connect their agents to Mubit
• Make infrastructure decisions that will still be the right ones at 100x scale

• You’ve built and run production systems — not just prototyped them
• You understand distributed systems, storage trade‑offs, and what it means to serve memory inside a hot execution loop
• You can go from a vague problem statement to a working system with minimal direction
• You have strong opinions about how infrastructure should be built — and you’re willing to defend them
• You care about the developer experience of the SDK you’re building — you’ve been the frustrated user before
• You’re based in London and want to work on‑site with the founding team, every day

• You need a well‑defined spec before you can start
• You’ve only ever built greenfield systems at scale — you’ve never had to make something work from nothing
• You want prototype forever and have someone else own production

Meaningful early‑stage equity — priced to reflect that you’re joining at the beginning.

Competitive salary — we’ll talk numbers early, no games.

Brand new Mac Book + hardware of your choice.

Private health insurance.

Canary Wharf office — flexibility to work from home when needed.

Deliveroo dinners when you’re working late.

Uber home if you’re in the office late — no one takes the night bus on our watch.