Machine Learning Engineer (Voice)

Salary: £100,000 - 100,000 per year

Requirements:

• We are looking for a cracked, AI-native engineer.
• You use AI to multiply what you ship and have made yourself measurably faster with coding agents or your own setup.
• You have shipped something that runs on an LLM in production.
• Voice AI experience is a big plus, including real-time STT, TTS, speech-to-speech, or telephony.
• You can show evidence of going above and beyond, such as a side project with users, a repo people star, a hackathon win, or something you built over a weekend.
• You have high agency and take ownership when something is broken.
• You prefer to move fast rather than wait for a spec.
• You have product taste and care how things look and feel, not just whether they work.
• You are low ego, kind, and hard-working.
• You care about the mission of helping older adults age well at home.
• Overlap with London hours matters more than where you live.

Responsibilities:

• Help turn our company into an intelligent system.
• Build speech-to-speech agents that feel like a person in the room.
• Work on interruptions, back-channels, noisy environments, hearing aids, thick accents, and millisecond latency.
• Build companions with long-term memory that recognize someone after six months like a friend would.
• Work on assistants inside messaging that connect families across cities and countries.
• Contribute to health biomarkers from conversation, including cognitive decline, mood, breathing, and loneliness.
• Work with research partners at Wharton, Toronto, and Oxford.
• Develop PhD-level audio in production, including speech enhancement, VAD, diarization, and TTS tuned for older voices.
• Build an agentic browser for booking appointments and getting things done online through conversation.
• Set engineering standards for testing, review, deployment, and the use of agents as we grow.
• Shape the stack and challenge how we build with strong evidence-backed opinions.
• Help interview and decide who joins our team.
• Own broadly across voice, memory, messaging, and biomarkers, and take primary ownership of a major surface as we grow.

Technologies:

• AI
• Support
• LLM

More:

We are Harmonica, an AI-native engineering organization building agents across coding, reviewing, QA, ops, support, scheduling, and more. We are focused on helping older adults age well at home through voice-first products, long-term memory, family messaging, health biomarkers from conversation, and an agentic browser that gets things done online. We work on production-grade audio and collaborate with research partners at Wharton, Toronto, and Oxford. This is a founding engineer role with real equity, a real voice in product, engineering, and company decisions, and the freedom to help shape the stack, the standards, and the team. We value mission, speed, ownership, product taste, and a low-ego, no-politics culture, and we care more about what you have built than where you studied or where you live.

last updated 38 week of 2026