CodeSignal Expands Cosmo's Agentic Capabilities — Any Hiring Team Can Now Create Expert-Level Skills Assessments in Minutes Using Talent Science AI
AI / Machine Learning | 3 min read
CodeSignal (San Francisco; the leading AI-native skills platform; Co-Founder and CEO Tigran Sloyan; millions of skills assessments completed; trusted by Netflix, Capital One, Meta, and Dropbox), has announced that any hiring team can now discover or create expert-level skills assessments and interview content on demand using Cosmo — CodeSignal's AI agent. Cosmo draws on tens of thousands of hours of talent science research produced by CodeSignal's Talent Science team — a group of Industrial-Organisational (I-O) psychologists who design and validate job-relevant assessments — to create simulation-based, role-specific assessments and interview content instantly for any role. The expansion builds on CodeSignal's April 2026 launch of industry-first agentic coding assessments — a new category of technical evaluation that measures what engineers can actually build when working with agentic AI tools.
"Building high-quality skills assessments and interview questions is hard. Our team of I-O psychologists develops the science behind what makes a great skills assessment, and thanks to AI, we're now delivering that expertise at scale and speed never seen before."
— Tigran Sloyan, Co-Founder and Chief Executive Officer, CodeSignal
How Cosmo Works — Discovery, Custom Creation, and Agentic Coding Assessments
Cosmo operates across two modes. In discovery mode, hiring teams can search CodeSignal's existing library of validated assessments — built by the Talent Science team — to find the closest match to their role requirements without starting from scratch. In custom creation mode, when no existing assessment fits precisely, Cosmo builds assessments, AI interviews, or live interview questions of all types from scratch — with the ability for hiring teams to give feedback and refine the output iteratively. The result mirrors the rigorous I-O psychology framework that CodeSignal's talent scientists apply manually — but delivered in minutes rather than weeks. The underlying science is central to CodeSignal's differentiation: assessments created by Cosmo are grounded in the same fairness standards, compliance frameworks, and job-relevance criteria as those built by the Talent Science team directly — giving hiring teams confidence that the content they generate is defensible, validated, and compliant. Alongside this expansion, CodeSignal's agentic coding assessments — launched 2 April 2026 — allow candidates to use tools like Claude Code, Cursor, and Codex to solve real-world engineering problems, giving employers a direct view of how candidates perform in the AI-native environments their teams actually use.
The AI Skill Shift — 91% of Engineers Already Use Agentic AI Tools at Work
The urgency behind CodeSignal's agentic assessment expansion is driven by a profound shift in how engineering work is actually done. A CodeSignal survey of 450 US software engineers conducted in March 2026 found that 91% already use agentic AI coding tools (such as Claude Code, Cursor, and Codex) in their work, and 75% have shipped production code partially or primarily generated with AI in the last six months. Yet until CodeSignal's agentic coding assessments, no hiring platform had reflected this reality in how candidates are evaluated. The data also underscores the competitive stakes: 73% of engineers believe those who do not adopt agentic AI tools risk becoming less competitive, with 30% saying that shift will come within the next year. And 56% said they would be hesitant to hire or work with an engineer who does not know how to use agentic AI coding tools. Across non-technical roles, Cosmo's expansion brings the same I-O-psychology-grounded assessment capability to sales, marketing, finance, HR, and operations — broadening CodeSignal's platform beyond engineering hiring into the full workforce.
Key Takeaways
- • CodeSignal (San Francisco; leading AI-native skills platform; Co-Founder and CEO Tigran Sloyan; millions of skills assessments completed; Netflix/Capital One/Meta/Dropbox customers) has announced expanded agentic capabilities for Cosmo — its AI agent — announced 28 April 2026. Any hiring team can now discover or create expert-level skills assessments and interview content on demand in minutes. Cosmo draws on tens of thousands of hours of talent science research from CodeSignal's Talent Science team (I-O psychologists) to create simulation-based, role-specific assessments and interview content for any role instantly.
- • Cosmo's two capabilities: (1) Discovery mode — search CodeSignal's existing validated assessment library to find the closest match for a role without building from scratch. (2) Custom creation mode — Cosmo builds assessments, AI interviews, or live interview questions of any type from scratch, with iterative feedback and refinement. All content grounded in the same I-O psychology fairness standards, compliance frameworks, and job-relevance criteria as assessments built by the Talent Science team directly — defensible, validated, and compliant.
- • Agentic coding assessments (launched 2 April 2026): a new category of technical hiring evaluation measuring what engineers can build using agentic AI tools including Claude Code, Cursor, and Codex. Candidates extract and interpret product/technical specifications, deploy agentic AI tools to devise functional solutions, and articulate their technical decisions to a human evaluator. CodeSignal survey of 450 US software engineers (March 2026): 91% use agentic AI coding tools at work; 75% shipped production code partially or primarily AI-generated in last six months; tens of thousands of candidates have completed AI-assisted assessments; one-third of CodeSignal customers adopted AI-assisted format in 2025.
- • The competitive stakes for AI skill adoption: 73% of engineers believe non-adopters of agentic AI tools risk becoming less competitive (30% say within the next year). 56% would be hesitant to hire or work with an engineer who does not know how to use agentic AI coding tools. Hiring hasn't reflected this reality — until CodeSignal's agentic assessments. The platform also measures real job skills across non-technical business roles: sales, marketing, finance, HR, and operations — making Cosmo's expansion relevant beyond engineering hiring to the full workforce.
- • Strategic significance: CodeSignal's Cosmo expansion represents the convergence of two trends simultaneously reshaping hiring — AI agents replacing manual expert labour (I-O psychologists building assessments) and AI tools becoming prerequisites for modern work (agentic coding skills as a baseline engineering requirement). By embedding I-O psychology research into an AI agent, CodeSignal makes expert-level assessment design accessible to any hiring team without assessment expertise. By launching agentic coding assessments, it becomes the first hiring platform where the evaluation mirrors how engineers actually work. The result: a hiring platform that is simultaneously more scientific (I-O-grounded), more current (agentic tool usage evaluated), and more accessible (any team in minutes) than anything that preceded it.