Senior Software Engineer, Data Acquisition

Who’re we?Our mission is to scale intelligence to serve humanity. We’re coaching and deploying frontier fashions for builders and enterprises who’re constructing AI programs to energy magical experiences like semantic search, RAG, and brokers. We consider that our work is instrumental to the widespread adoption of AI.
We obsess over what we construct. Every one in every of us is accountable for contributing to rising the capabilities of our fashions and the worth they drive for our clients. We wish to work exhausting and transfer quick to do what’s finest for our clients.
Cohere is a group of researchers, engineers, designers, and extra, who’re obsessed with their craft. Every individual is the probably the greatest on the planet at what they do. We consider {that a} various vary of views is a requirement for constructing nice merchandise.
Be a part of us on our mission and form the longer term!
Why this function? At Cohere, we try to repeatedly enhance our giant language fashions. Educational analysis and real-world expertise has demonstrated that top high quality, various datasets can contribute as a lot to the efficiency and capabilities of LLMs because the underlying mannequin structure and coaching routine. We at Cohere consider knowledge will play a central function in accelerating the development of our already world-class language fashions.
Information is due to this fact vital to our success. Our potential to accumulate knowledge that’s correct, related, and well timed is essential to our potential to enhance the standard of our fashions. We try to repeatedly enhance our knowledge acquisition processes and programs to make sure that we’ve the information we have to keep aggressive and meet the wants of our clients. We run frequent experiments to be taught extra concerning the function of information for mannequin high quality, from knowledge mixtures, to cleansing strategies, to high quality management.
This function can be a part of the Information Acquisition group, which broadly supplies knowledge for coaching fashions and is accountable for constructing and sustaining the infrastructure that acquires, cleans, and codecs knowledge for mannequin coaching. We’re searching for a technically expert, resourceful problem-solver who is ready to work in areas of ambiguity and discover environment friendly and generally inventive options. The primary accountability of this function is to enhance our inside knowledge acquisition infrastructure, which incorporates knowledge crawlers, formatters, and integrations with knowledge suppliers. This function would additionally work intently with totally different groups at Cohere to help their knowledge acquisition wants, in addition to have interaction in additional experimental work to develop extremely informative knowledge indicators.
Please Word: We have now places of work in Toronto, Palo Alto, and London however embrace being remote-first! There are not any restrictions on the place you might be positioned for this function.

As a Senior Software program Engineer specializing in Information Acquisition, you’ll:

• Purchase, put together, and combine high-quality datasets into mannequin coaching and analysis
• Design, construct, and optimize large-scale knowledge pipelines for ingesting and processing knowledge from various sources
• Collaborate with analysis and product groups to determine, prioritize, and safe new knowledge sources
• Develop tooling and infrastructure to systematically consider and distill giant datasets
• Mentor junior group members and contribute to their skilled progress

You might be match if:

• You could have 5+ years of expertise in software program improvement, knowledge engineering, or machine studying
• You could have written, debugged, and optimized Spark code over petabyte-scale datasets
• You could have proficiency in Python
• You could have efficiently carried out knowledge high quality options in manufacturing
• You might be aware of orchestration engines like Airflow, Dagster, and Prefect
• You could have skilled classifiers to carry out large-scale knowledge labelling
• You might be comfy utilizing embeddings or language fashions as a part of your workflow
• You could have robust communication and problem-solving abilities, and like utilizing the precise device for the job even when it’s outdoors your wheelhouse
• You could have a demonstrated ardour for utilized NLP fashions and merchandise

If a few of the above doesn’t line up completely along with your expertise, we nonetheless encourage you to use! If you happen to contemplate your self a considerate employee, a lifelong learner, and a form and playful group member, Cohere is the place for you.
We worth and have a good time variety and attempt to create an inclusive work surroundings for all. We welcome candidates of all types and are dedicated to offering an equal alternative course of. Cohere supplies accessibility lodging through the recruitment course of. Must you require any lodging, please tell us and we’ll work with you to satisfy your wants.
Our Perks: An open and inclusive tradition and work surroundings  ‍ Work intently with a group on the slicing fringe of AI analysis  Weekly lunch stipend, in-office lunches & snacks Full well being and dental advantages, together with a separate price range to care for your psychological well being  100% Parental Depart top-up for six months for workers primarily based in Canada, the US, and the UK Private enrichment advantages in direction of arts and tradition, health and well-being, high quality time, and workspace enchancment Distant-flexible, places of work in Toronto, Palo Alto, San-Francisco and London and co-working stipend✈️ 6 weeks of trip
Word: This put up is co-authored by each Cohere people and Cohere know-how.