Roster

Roster uses machine learning to match creator job postings with freelancer profiles by analyzing portfolio artifacts (videos, design files, audio samples), work history, and skill tags to infer creative competencies. The system likely employs embeddings-based similarity matching or collaborative filtering to rank talent candidates by relevance to specific creative roles (motion designer, colorist, sound engineer), reducing manual screening time for creators unfamiliar with evaluating technical creative work.

Roster implements a vetting pipeline to validate freelancer credentials, work samples, and past project quality before surfacing them to creators. This likely includes portfolio authenticity checks (verifying work samples are genuinely the freelancer's), skill validation through past client feedback or test projects, and possibly credential verification for specialized roles. The system maintains a curated talent pool rather than open-marketplace model, reducing creator friction from low-quality or fraudulent profiles.

Roster provides a freemium job posting interface where creators can describe projects, required skills, and budget without payment friction. The discovery layer allows browsing vetted freelancer profiles filtered by specialization (video, design, audio), experience level, and past work. This combines traditional job-board functionality with portfolio-first discovery, enabling creators to explore talent before committing to hiring or premium features.

Roster maintains a specialized taxonomy of creative roles (motion designer, colorist, sound engineer, video editor, etc.) and associated skill tags, enabling precise filtering and matching. The system likely maps freelancer profiles and job postings to this taxonomy, allowing creators to filter talent by specific creative specializations rather than generic job titles. This domain-specific structure enables more accurate matching and discovery than generalist freelance platforms.

Roster analyzes freelancer portfolio artifacts (video files, design images, audio samples) to infer creative skills and quality without relying solely on text descriptions or self-reported tags. This likely involves computer vision (analyzing video frames for color grading, motion design complexity, visual effects quality) and audio analysis (evaluating sound design, mixing quality) to validate claimed skills. The system may extract metadata from portfolio files (software used, project complexity) to enrich freelancer profiles.

Roster provides in-platform messaging and project coordination tools enabling creators to communicate with matched or discovered freelancers, negotiate terms, and manage project scope. The system likely includes contract templates, milestone tracking, and file sharing to streamline the hiring-to-delivery workflow. This reduces friction of moving conversations off-platform (email, Slack) and enables Roster to track project outcomes for matching algorithm feedback.

Roster implements a structured onboarding flow for freelancers to create profiles, upload portfolio samples, and complete skill assessments or vetting questionnaires. The system likely guides freelancers through portfolio upload (video, design, audio files), skill tag selection, rate setting, and availability scheduling. This standardized onboarding ensures profile completeness for matching and vetting, reducing friction for freelancers unfamiliar with portfolio-first platforms.

Roster implements a freemium model where creators can post jobs and browse talent without payment, with premium features (likely enhanced matching, priority support, advanced filtering, or direct messaging) behind a paywall. The system tracks creator engagement (job postings, talent browsing, hires) to identify conversion opportunities and optimize pricing. This model reduces friction for bootstrapped creators while generating revenue from successful hires or feature upgrades.