Senior AI Architect & Founding Member – Real-Time Translation

We’re building the world’s first real-time, bidirectional translation for live conversations. No language barriers left—just seamless, human-level conversations in 250+ languages, as easily as old friends already do.

That’s not what exists yet. ✨🌍 We’re fixing it.

Meet the Team

We’re a small but ambitious team in vibrant London, backed by industry leaders and veterans of tech we actually use every day. Right now, the technology doesn’t exist—and that’s the exciting part.

This is founding-level in every sense.

The Role

You’ll join our AI leadership committee as an indispensable member (seat at the table alongside the CEO, Head of Engineering, Head of Product, Head of Growth & Partnerships, and Chief of Staff). Your expertise will define:

• AI strategy and architectural direction
• Model architecture & long-term technical bets
• How the product thinks, learns, scales and interacts organically with the end user
• The technical principles that will drive us forward for years

You won’t be a typical manager or vice-president—this role belongs to someone who speaks at the same level as the buy-in decisions and helps telescope our vision directly.

Core Responsibilities

• Shape the technical direction of conversation AI, from model architecture to deployment pipelines.
• Translate abstract research into product-ready prototypes, ensuring graceful trade-offs between accuracy, latency and scalability.
• Solve genuinely hard, unsolved problems that no one—including us—has tackled effectively before (low-latency speech processing, cross-lingual alignment, live-inference scaling).
• Guide engineeringTeams in best practices for distributed machine learning, spoken language systems and dynamic reranking in real time.
• Champion_clean coding, rigor, and replicability, but don’t stop to *ask fork a mutable state over deterministic alternatives—just ship the solution you’re confident in.
• Translate grandiose goals into Webhooks or rate-limiters. Your execution should flow from both technical integrity and user impact.
• Build thrRust.

Expected Background

You bring:

• Deep end-to-end AI engineering experience—coding to fine-tuning, architecture to systems running in production at scale.
• Hard engineering challenges you’ve solved in your bag. Real-time processing, speech-to-speech translation (e.g., Whisper), hetero-genization of inference, low-latency cascade models, or highly available, low-latency pipelines (perhaps serving the frontline of Google Voice or prototype piloted at DeepMind).
• Proven judgement at the roulette table: robust whole-program decisions that you’re trusted to commit to (and slide carefully on losses). People trusted to say “no, we’ll stick to this now,” then deliver on why it’ll pay dividends later.
• Regligious about end-to-end engineering: pushing our models to class-leading regimes in their domains (it’s obvious when “the interns will build this later” turns out to be 10 weeks of contrived CLI PL/Python GU acrophonics scavenged from deprecated frameworks).
• Good enough coding ability to cut your own trail. Zero need to throw all dependentials into metapkg pitholes. You understand the gigantic 1-liner versus CG thread bailout tradeoff, and you aim for do-ones love minimally.
• You skip the allure of shiny toys. You’re here to solve ambiguity.
• "You will be building relationships }, as many of our partners have You will be superlative to the core bot.
• A statement that reflects on where you are in the present and how you plan to shape the future of this world.