Senior Research Engineer (LLM Post-Training £300k)

{ “@context”: “http://schema.org”, “@type”: “JobPosting”, “title”: “Senior Research Engineer (LLM Post-Training £300k)”, “description”: “

Location: London, 5 days in office

Salary: Up to £300,000 + Equity

The role

Senior research engineer experienced in post‑training large language models, with a focus on scaling, RLHF, DPO, reward modelling, and fine‑tuning built on strong base models. The team is small, the bar is high, and the work ships into product used by enterprise customers.

The work

  • Run post‑training pipelines and own model improvements end‑to‑end
  • Build evaluation frameworks that distinguish real capability gains from benchmark theatre
  • Debug training runs, reason about model behaviour, ship better weights
  • Work directly with product on what to train for next

What you bring

  • Direct experience post‑training or fine‑tuning LLMs at scale
  • Strong PyTorch and distributed training fundamentals
  • Published research, serious open‑source, or shipped production model work

#J-18808-Ljbffr”, “datePosted”: “2026-05-07”, “hiringOrganization”: { “@type”: “Organization”, “name”: “Dex”, “sameAs”: “https://uk.whatjobs.com/pub_api__cpl__425535455__4861?utm_campaign=publisher&utm_medium=api&utm_source=4861&geoID=33” }, “jobLocation”: { “@type”: “Place”, “address”: { “@type”: “PostalAddress”, “addressLocality”: “London” } } }
Company: Dex
Apply for the Senior Research Engineer (LLM Post-Training £300k)
Location: London
Job Description:

Location: London, 5 days in office

Salary: Up to £300,000 + Equity

The role

Senior research engineer experienced in post‑training large language models, with a focus on scaling, RLHF, DPO, reward modelling, and fine‑tuning built on strong base models. The team is small, the bar is high, and the work ships into product used by enterprise customers.

The work

  • Run post‑training pipelines and own model improvements end‑to‑end
  • Build evaluation frameworks that distinguish real capability gains from benchmark theatre
  • Debug training runs, reason about model behaviour, ship better weights
  • Work directly with product on what to train for next

What you bring

  • Direct experience post‑training or fine‑tuning LLMs at scale
  • Strong PyTorch and distributed training fundamentals
  • Published research, serious open‑source, or shipped production model work

#J-18808-Ljbffr…

Posted: May 7th, 2026