Job Description
Perplexity is offering a machine learning research engineer internship in Berlin, running 12 to 24 weeks full-time in the office, to improve search quality through retrieval and ranking models, representation learning research and retrieval-augmented generation pipelines.
At a glance
- Role: Internship, Machine Learning Research Engineer
- Where: Berlin, Germany (in person)
- Length: 12–24 weeks, full-time
- Focus: retrieval, ranking and representation learning for AI search
- Apply: open until filled (re-checked 12 October 2026)
About the ML research internship
- Push search quality forward through models, data, tools or any other lever available
- Train and optimise large deep learning models in PyTorch using distributed training (such as PyTorch Distributed, DeepSpeed or FSDP), with a focus on retrieval and ranking
- Research representation learning, including contrastive, multilingual and multimodal modelling and evaluation for search
- Build and optimise RAG pipelines for grounding and answer generation
Modern AI search depends on retrieval and ranking models that decide which documents an answer is built from. Better representations, especially across languages and media, make answers more accurate and better grounded, which is why this internship combines research with hands-on model training.
What Perplexity is looking for
- Understanding of search and retrieval systems, including quality evaluation principles and metrics
- Strong PyTorch, with experience in distributed training and performance optimisation for large models
- Interest in representation learning: contrastive learning, dense and sparse representations, fusion, cross-lingual alignment, training data optimisation and robust evaluation
- A publication record at AI/ML conferences or workshops such as NeurIPS, ICML, ICLR, ACL, EMNLP or SIGIR
Pay and location
Perplexity’s posting does not state intern pay. The internship is in person in Berlin for 12 to 24 weeks.
How to apply
Apply through the official Perplexity job posting. Perplexity’s posting does not give a closing date, so the role is open until filled; ResearchJobs.in will re-check this listing on 12 October 2026. Check the posting for location and work-authorisation details before you apply.
See all our artificial intelligence jobs, or browse more research jobs on ResearchJobs.in.
Hiring institution: Perplexity
Official advertisement: jobs.ashbyhq.com
How to prepare for this application
- Retrieval metrics: revise NDCG, MRR and recall@k, and their limits for AI search.
- Contrastive learning: be ready to explain hard-negative mining and in-batch negatives.
- Multilingual retrieval: think about cross-lingual alignment and evaluation.
- Distributed training: know FSDP and DeepSpeed basics.
- Publications: the posting asks for a publication record, so lead with your strongest paper.
About Perplexity
Perplexity builds AI-powered search and agent products, including its Sonar models, Deep Research Agent, Comet Agent and Search, which handle hundreds of millions of queries and are growing rapidly.