Staff ML Research Engineer | LLM Fine-Tuning | RL | SFT | DPO | Must Have Startup Experience

Salary: £210,000 - £305,000 + meaningful equity

Contract: Permanent Start: ASAP

Working model: Full-time onsite (some flex around this) - Shoreditch, London

Eligibility: Must have existing UK work authorisation - Dependent visa is fine

What You'll Get

• £210,000 – £305,000 + meaningful equity
• Full medical, dental, and vision coverage
• Uncapped holiday - take what you need, when you need it
• Daily meals covered - breakfast, lunch, dinner, and snacks on site every day

Optimal has teamed up with an exciting AI company based in London that’s growing its research team, offering a rare opportunity to join at just the right time.

This is a company that has built its own proprietary models, has serious commercial traction, and is scaling fast. They're not wrapping third-party APIs and calling it AI. They've done the hard work - and it's paying off.

The research function sits at the core of everything. You'll be building AI systems that push the boundary of what's possible - designing and implementing state-of-the-art methods for instruction tuning,

information retrieval, and agent capability. Your work won't sit in a repo waiting for review. It'll ship, it'll scale, and it'll be used by real customers.

This is a small, elite team. Everyone here has high ownership, high impact, and zero time for passengers. If that sounds like your environment - keep reading.

Please only apply if you have ALL of the following

• 5+ years in applied AI/ML engineering or research (exceptional candidates at 3+ considered)
• Hands-on, production-grade experience fine-tuning LLMs (SFT, DPO, PPO, GRPO, RL)
• Proven track record deploying models in live, customer-facing environments
• Experience working with large open-source LLMs (e.g. Llama or similar)
• Startup or scale-up experience - you've moved fast, owned outcomes, shipped real things
• Strong software engineering fundamentals - this is not a data science or research-only role

ℹ️ Very Important Notes

• This role requires deep applied ML engineering - not suitable for pure researchers or academic profiles
• You must be comfortable in a fast-moving, high-ownership startup environment
• You'll be expected to move from research idea → production system at pace
• Onsite in Shoreditch, London - Fulltime (staff-level candidates have some flexibility)

Must-Haves

• Production-grade LLM fine-tuning experience - SFT, RL, DPO, PPO, GRPO
• Deep familiarity with large open-source language models
• Strong software engineering skills - you write clean, bug-free ML code
• Ability to break down ambiguous research problems into clear, shippable milestones
• Startup mindset - evidence of high personal achievement and genuine ownership
• Excited about applied product impact, not just foundation model or academic research

Bonus Experience

• Voice and speech model experience
• Multimodal model exposure
• Experience generating synthetic data and building evaluation pipelines
• Background from product-driven orgs (not just research labs)
• Prior founding engineer or early-stage startup experience

What You'll Be Doing

Research & Model Development

• Build and fine-tune models for complex, real-world customer-facing tasks
• Run experiments with open-source LLMs to drive order-of-magnitude gains in latency and performance
• Design and implement state-of-the-art instruction tuning and information retrieval methods

Production Ownership

• Take research from prototype to fully deployed, production-grade system
• Validate model behaviour against real-world workflows and user feedback
• Improve reliability, capability, and performance of live AI systems

Collaboration

• Work directly alongside science and engineering teams on new architectures
• Feed real-world findings back into platform evolution and roadmap
• Break down research ideas into clear, iterative milestones

What They're Looking For

• A technically sharp, applied engineer with a founder-like mentality
• Someone who thrives in ambiguity and moves with urgency
• A builder who wants to ship things that matter, not just publish papers
• An engineer who's energised by ownership, pace, and real-world impact

If you meet the requirements above and want to do some of the most impactful ML research work in London right now - get in touch for a fast response.