Find a job 
Featured Companies 
Featured Universities 
For Companies At a Glance
- Tasks: Work on cutting-edge AI models and see your impact in real-world applications.
- Company: Innovative tech firm at the forefront of AI research.
- Benefits: Competitive salary, dynamic work environment, and opportunities for professional growth.
- Other info: Collaborate with high-calibre peers in a fast-paced, impactful role.
- Why this job: Join a small team where your contributions directly influence live AI systems.
- Qualifications: 5+ years in ML engineering with hands-on production experience.
Most AI research roles sit at one of two extremes - pure academia with little real-world impact, or pure engineering with little room to think. This one sits in the narrow band between them, where the work is serious, the models are live, and the problems are genuinely unsolved. You'll work on the core model stack - post-training, fine-tuning, alignment, evaluation - and see the results in production within weeks, not quarters. The measure of success is model performance in the real world, not benchmark scores or citation counts. Small team. High ownership. High-calibre peers.
WHAT YOU'LL WORK ON
- Post-training & alignment: Designing and running SFT pipelines and applying alignment techniques - RLHF via PPO or the more direct DPO family. You understand what each approach optimises for, where each breaks down, and which to reach for given the data and objective at hand.
- Parameter-efficient fine-tuning: Production experience with LoRA and QLoRA. You understand how rank, the alpha-to-rank scaling ratio, and target module selection interact with model behaviour.
- Evaluation & failure diagnosis: Building eval frameworks tied to real-world outcomes. You can identify why a model is failing from first principles.
- Training infrastructure: Owning data pipelines, distributed training runs, and model versioning end-to-end. PyTorch is your default. You've debugged a training run that wasn't converging and knew where to look.
WHAT WE NEED TO SEE
- You have taken an LLM through post-training and into a live production environment.
- You can describe a specific model behaviour you improved, what you changed, and why it worked.
- 5+ years in ML engineering or applied research with a clear production track record.
- Deep Python across data, training, evaluation and serving - no significant gaps.
- Strong academic backgrounds, provided they come with at least two years of hands-on production ML experience.
Research Technology Engineer in London employer: DURLSTON PARTNERS
Contact Detail:
DURLSTON PARTNERS Recruiting Team
StudySmarter Expert Advice 🤫
We think this is how you could land Research Technology Engineer in London
✨Tip Number 1
Network like a pro! Reach out to folks in the industry on LinkedIn or at meetups. We all know that sometimes it’s not just what you know, but who you know that can land you that dream job.
✨Tip Number 2
Prepare for those interviews by brushing up on your technical skills and problem-solving abilities. We recommend doing mock interviews with friends or using platforms that simulate real interview scenarios. Practice makes perfect!
✨Tip Number 3
Showcase your projects! If you've worked on any ML models or research, make sure to have a portfolio ready. We love seeing real-world applications of your skills, so don’t hold back on sharing your successes.
✨Tip Number 4
Apply through our website! It’s the best way to ensure your application gets seen by the right people. Plus, we’re always on the lookout for passionate candidates who want to make an impact in the AI space.
We think you need these skills to ace Research Technology Engineer in London
Some tips for your application 🫡
Show Your Passion for AI: When you're writing your application, let your enthusiasm for AI and machine learning shine through. We want to see that you’re not just ticking boxes but genuinely excited about the challenges and innovations in this field.
Be Specific About Your Experience: Don’t just list your past roles; dive into the details! Share specific examples of projects where you've taken models from post-training to production. We love seeing how you’ve tackled real-world problems and what impact your work had.
Tailor Your Application: Make sure your application speaks directly to the job description. Highlight your experience with SFT pipelines, fine-tuning techniques, and evaluation frameworks. We appreciate when candidates connect their skills to what we’re looking for!
Keep It Professional Yet Personal: While we love a friendly tone, remember to keep it professional. Share a bit about yourself and why you want to join us at StudySmarter, but also ensure your application reflects your expertise and seriousness about the role.
How to prepare for a job interview at DURLSTON PARTNERS
✨Know Your Models Inside Out
Make sure you can discuss the models you've worked with in detail. Be prepared to explain your experience with post-training processes, fine-tuning techniques like LoRA and QLoRA, and how you've applied alignment methods. This shows you not only understand the theory but also have practical experience.
✨Demonstrate Real-World Impact
Since this role focuses on real-world outcomes, come ready with examples of how your work has directly influenced model performance in production. Share specific instances where you diagnosed failures or improved model behaviour, and be clear about the changes you made and their results.
✨Showcase Your Problem-Solving Skills
Prepare to discuss challenges you've faced in ML engineering, particularly around training infrastructure and debugging. Highlight your thought process when diagnosing issues with training runs and how you approached finding solutions. This will demonstrate your critical thinking and hands-on expertise.
✨Be Ready for Technical Questions
Expect deep technical questions related to Python, data pipelines, and distributed training. Brush up on your knowledge of these areas and be ready to solve problems on the spot. Practising coding challenges or discussing past projects can help you feel more confident during this part of the interview.