Cognition jobs, judged for you, free

What they do

Cognition is an applied AI lab that builds end-to-end software agents. It makes Devin, which it calls the first AI software engineer, and Windsurf, an AI-native IDE (a code editor). Devin plans, writes, tests, and ships production code inside a team's own codebase and tools. Cognition says its revenue has passed a pace of $1B a year.

What they value

• Urgency: the team says it is racing to build the future of software engineering.
• A small, talent-dense team, with former founders and top competitive programmers.
• Teamwork and in-person work: it says its best work is done in person.
• Technology should expand what people can do, working beside them rather than replacing meaningful work.

Where they hire

San Francisco has the most roles, then London, New York, Singapore, Tokyo, Washington, DC, and Austin. Others are in Sydney, India, Riyadh, São Paulo, Doha, South Korea, Mexico City, Berlin, and Southern Europe. Most roles are on site, and some say five days a week. Sydney and Riyadh roles are remote. No Posting mentions visa sponsorship.

Teams and roles

The 103 Postings split across Sales (30), General and Administrative (27), Customer Engineering (24), Marketing (13), and Research and Development (9).

• Customer engineering: deployed engineers, who put Devin to work for customers, partner deployed engineers, applied AI engineers, and AI support engineers who work with customers in many regions, including a federal team in Washington, DC.
• Sales: account directors by region, sales leaders, and sales development representatives.
• Research and development: research engineers who train its AI models (mid-training and post-training) and build the systems for that training, plus software, infrastructure, security, data, and site reliability engineers, who keep systems running.
• Marketing: brand, events, paid media, product marketing, and developer community.
• General and administrative: recruiting, people operations, IT, legal, payroll, and workplace roles.

What they look for

Many Postings list requirements and then "you might excel if you". Software engineers need distributed systems depth, product instinct, and strong Python. The post-training research engineer needs a record with methods such as RLHF, which trains a model from human feedback, and strong statistics. Deployed engineers need a strong engineering base, a record of top performance in any field, the drive to work without a manager saying what to do next, and travel of 25 to 50 percent. Applied AI engineers need 3+ years in a coding role and a degree in science, technology, engineering, or maths (STEM), or equal experience. AI support engineers need 4+ years and Linux, Docker, Git, and cloud skills. Account directors need a record of selling technical products to enterprises. Sales development representatives need 1+ years in sales or success in another field. Japan roles ask for fluent Japanese and English. Many sales and deployed engineer roles list founding a startup as a plus. Many roles ask for comfort with ambiguity and fast change.