Winter 2026
Machine Learning Intern/Co-op
Winter 2026
Posted on 10/31/2025
Cohere
501-1,000 employees
API-based NLP tools and LLMs
No salary listed
London, UK + 3 more
More locations: Toronto, ON, Canada | San Francisco, CA, USA | New York, NY, USA
Remote
Hybrid role; some on-site days at Toronto, New York, San Francisco, or London offices.
 AI & Machine Learning (1) |
|---|
- Student currently enrolled in a post-secondary program, available for a full-time 3-6 month internship, co-op, or research work term
- Design, train and improve upon cutting-edge models.
- Help us develop new techniques to train and serve models safer, better, and faster.
- Train extremely large-scale models on massive datasets.
- Explore continual and active learning strategies for streaming data.
- Learn from experienced senior machine learning technical staff.
- Work closely with product teams to develop solutions.
- Experience writing kernels for GPUs using CUDA.
- Experience training on Tensor Processing Units.
- Papers at top-tier venues (such as NeurIPS, ICML, ICLR, AIStats, MLSys, JMLR, AAAI, Nature, COLING, ACL, EMNLP).
- Proficiency in Python and related ML frameworks such as Tensorflow, TF-Serving, JAX, and XLA/MLIR.
- Experience using large-scale distributed training strategies.
- Familiarity with autoregressive sequence models, such as Transformers.
- Strong communication and problem-solving skills.
- Demonstrated passion for applied Natural Language Processing models and products.
Cohere provides access to advanced Natural Language Processing (NLP) tools and Large Language Models (LLMs) through a simple API. It serves businesses that want to improve content generation, summarization, and semantic search across multiple languages. The product works by offering API access to pre-trained models that perform tasks like text classification, sentiment analysis, and semantic search; users can customize and integrate these models into their applications, enabling scalable and affordable AI-powered solutions. Cohere differentiates itself with a developer-friendly API, multilingual support, and easy customization to help organizations build smarter and faster AI solutions. The company’s goal is to make powerful generative AI tools accessible to a wide range of customers and use cases, letting them deploy AI features quickly without managing complex models themselves.
Company Size
501-1,000
Company Stage
Series E
Total Funding
$2.1B
Headquarters
Toronto, Canada
Founded
2019
Simplify's Take
What believers are saying
- Cohere tripled revenue past $150M in 2025 with clients RBC, BCE, and SAP.
- Schwarz Group invests $600M in Series E amid Aleph Alpha merger valuing $20B.
- Cohere partnered with Ensemble for first RCM-native LLM in healthcare.
What critics are saying
- Aleph Alpha merger causes cultural clashes, defecting clients to Oracle within 12 months.
- Tiny Aya open-sourcing cannibalizes premium API revenue to HuggingFace in 6 months.
- US scrutiny revokes FedRAMP High over merger, collapsing 25% ARR in 12 months.
What makes Cohere unique
- Cohere achieved FedRAMP High authorization in under 90 days via Second Front.
- Cohere launched North agentic AI platform for enterprise workplace productivity.
- Cohere offers zero data retention and opt-out training for enterprise customers.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Health Insurance
Dental Insurance
100% Parental Leave top-up
Weekly lunch stipend
Remote Work Options
6 weeks of vacation
Growth & Insights and Company News
6 month growth
↓ -2%1 year growth
↓ -3%2 year growth
↑ 0%Ensemble, a US revenue cycle management services provider, has partnered with enterprise AI company Cohere to build the healthcare industry's first revenue cycle management-native large language model. The companies are creating a custom model informed by Ensemble's operational expertise and data, designed to handle complex healthcare financial operations more accurately than general-purpose LLMs. The model will be embedded into AI agents managing processes from patient intake to account resolution. Unlike standard approaches that rely on prompt engineering, this system is fine-tuned on real RCM tasks and trained using synthetic datasets in a HIPAA-compliant environment, without using identifiable patient data. The solution aims to enhance existing electronic health record systems by providing better context and guidance for navigating payer requirements whilst reducing administrative burden for healthcare providers.
Cohere has launched Tiny Aya, a family of open-weight multilingual AI models supporting over 70 languages that can run on everyday devices without internet connectivity. The models were unveiled at the India AI Summit by the company's research arm, Cohere Labs. The base model contains 3.35 billion parameters and includes regional variants: TinyAya-Global for broad language support, TinyAya-Earth for African languages, TinyAya-Fire for South Asian languages, and TinyAya-Water for Asia Pacific, West Asia and Europe. South Asian language support includes Bengali, Hindi, Punjabi, Tamil and Telugu. Trained on 64 Nvidia H100 GPUs using modest computing resources, the models enable offline applications like translation, particularly useful in linguistically diverse countries like India. The models are available on HuggingFace, Kaggle and the Cohere Platform.
SAP and Cohere are expanding their partnership to deliver sovereign AI solutions globally, beginning in Canada. SAP Canada plans to integrate Cohere's agentic platform, North, into its Enterprise Resource Planning Sovereign Cloud environment, creating a complete Sovereign AI Layer for public sector and regulated industries. The integration embeds Cohere's large language models into SAP's Canadian-operated sovereign cloud infrastructure, allowing organisations to deploy advanced AI whilst maintaining data residency and operational control. This addresses the challenge of innovating with AI without compromising security or data sovereignty. A recent SAP AI report found that whilst 71% of organisations rely on data for investment decisions, 75% report incomplete data as a significant challenge. The partnership aims to overcome data fragmentation by embedding AI directly into core SAP applications.
Toronto-based Cohere raised $600 million in 2025, achieving a $7 billion valuation, as the generative AI company secured contracts with major clients including RBC, Bell, Dell, Thales, SAP and LG for its office automation software. The company, which hired researcher Joëlle Pineau as chief AI officer, entered 2025 with approximately $50 million in annualised revenues and exited the year at more than triple that level. Chief executive Aidan Gomez expects dramatic growth to continue in 2026. Cohere has joined an elite group of 77 Canadian technology companies surpassing $100 million in annual revenue, a key threshold for sector maturity. The company also expanded internationally, opening offices worldwide during its breakthrough year.
AI startup Cohere Inc. has secured $500 million in new funding, valuing the company at $6.8 billion. This funding round is part of Cohere's strategy to compete with larger tech firms.