Amazon Science
Applied Scientist II, Alexa for Shopping Science UK
How your CV stacks up
Upload your CV to see how well it fits this job role
?%
Applied Scientist (Language Technology for Alexa for Shopping)
We are looking for a passionate, talented, and inventive Applied Scientist with a strong machine learning (ML) background to help build industry-leading language technology powering Alexa for Shopping, our AI-driven search and shopping assistant, enabling seamless, intuitive shopping support for customers at every stage of their journey.
This innovative role centres on developing and optimising large language model (LLM)/small language model (SLM)-powered conversational experiences by leveraging:
- Careful instruction design
- Contextual grounding
- Seamless multi-agent/multi-tool system orchestration
- Advanced prompt engineering
- Model fine-tuning & inference optimisation
- Robust evaluation frameworks
- Scalable experimentation techniques
Our mission is to transform shopping by delivering the most personalised, efficient, and delightful product discovery and purchase process. This is achieved through:
- AI-driven product research & recommendations
- Comparative analytics & visual shopping assistance
- Instant answers to product inquiries
- Multimodal shopping interactions (integrating images, video, and multimodal AI)
- Advanced shopping automation (e.g., price alerts, hands-free purchases)
Proactively shaping the future of AI-driven shopping experiences, your role spans scientific rigour and cross-functional partnership, applying machine learning and neurosymbolic techniques to systematically boost conversational AI performance at enterprise scale.
Key Responsibilities
Core Technical Leadership
- Treat LLMs/SLMs as production-grade agents, designing and maintaining:
- Automated evaluation pipelines for nuanced feedback
- LLM-as-a-judge systems, rubrics, and custom datasets
- Advanced post-learning pipelining (e.g., in-context learning splits, robotic evaluation)
- Collaborate with engineers and researchers to explore LLM augmentation strategies:
- Retrieval-augmented generation for retrieval + reasoning convergence
- Contextual enrichment to debias & improve robustness
- Prompt decomposition for improved instructionfähigkeit
- Model optimisations (supervised fine-tuning, reinforcement learning, NLP distillation et al.)
- Optimise architectures for cost / latency consistency, including:
- Post-training for small, efficient models (SLMs, LLMs with strong latency constraints)
- Distributed fine-tuning and on-device interpretability
- Develop ragioning + visual modelling innovations, such as:
- Multimodal engagement (multicolors, invert checks, remote command targeting)
- Advanced 3D query scaffolding techniques
Reasons to use Rodeo
Start with a chat, not a search bar
Grad scheme, placement, apprenticeship? Not sure what you want yet — that's fine. Your agent talks it through with you and turns "I have no idea" into a shortlist.
It searches the market for you
Every day your agent scans the market matching roles against what actually matters to you, not just keywords on a CV.
Only hits
No noise. No "maybe this fits." Just roles with a clear explanation of why they're right — and where to focus when applying.
End-to-End AI Impact
- Conduct large-scale multimodal dataset analyses, inferring user pain points and conversational AI paradigms where shoppers falter
- Design A/B tests and experiments to empirically evaluate and iterate on:
- Conversational quality (MbS versus CAB evaluations)
- Latency-sensitive optimisations
- Page-grounding and personalization
- Translate insights into production 효율물론기 sequence:
- Meet with product managers and engineers
- Synthesize actionable design proposals & address ML questions
Organisational & Collaborative Impact
- Represent the team at scientific conferences & internal forums
- Author internal delivery reports and cross-group design documentation
- Present insights via engineering_info meetings, cultural workshop led topics and investor-facing summaries
About the Team
The Alexa for Shopping Science team based in London supports ~150 engineers, designers, and product managers, focusing on:
- Agentic design – Dynamic price alert interfaces, honoring a customer’s episodic classes, self-checkouts
- Multimodal conversational AI from API interface to context-grounded outputs
- Deep research prowess, analysing Amazon’s product catalogue and public sources to uncover personalised insights
- Cutting-edge development in NLP, generative AI, information retrieval, ML, and lifting biology
- State-of-the-art performance measures through rigorous empirical validation: publishing externally
Get help with your application
Your very own career expert that helps elevate your application to the next level.
Basic Qualifications
- A PhD in CS, CE, ML, data science (or equivalent master’s degree) with demonstrated research and pixelation track
- Experience designing and optimising deep learning architectures incl:
- LLMs / transformers
- Self-supervised / decoding architectures
- Recovery and distraction-based pruning
- Extensive patent filings / peer-reviewed publications in top-tier NLP/ML venues (e.g., NEURIPS, ICLR, ACL, NAACL, VLDB)
- Strong programming aptitude in Java, C++, or Python, with an encyclical grasp of parallel, distributed, and high-performance computing
- Proven success applying ML to product analytics, conversational AI, or data-driven workloads
Preferred Qualifications
- Deep expertise in LLM post-training (supervised fine-tuning, RLHF, rewards model iteration)
- Hands-on experience with large-scale distributed training pipelines
- Exemplary citation count & prestige publications: >500 citations from top-tier NLP/LLM conferences incl.:
- ICLR, NeurIPS, ACL, EMNLP, ICML, VMCAI, SIGIR
- Advanced inherent model explanation and alignment techniques
- Broad impact mentorship in internal teams and scientific communities
“It took my CV and asked me questions relevant to understanding what kind of jobs to suggest for me. Suggestions were almost perfect. Jobs were exactly what I’ve been looking for.”
Jessica, London
Skills
Location