Lever ATS Resume Guide: Format, Parsing, and Ranking Tips

Who Uses Lever and Why It Matters

Lever sits at the center of the modern tech hiring stack. Its customer base skews toward venture-backed startups, Series B-and-later growth companies, mid-market SaaS firms, and the technology divisions of larger enterprises that want a faster, more relationship-oriented recruiting workflow than Workday or Oracle Taleo can offer. The platform is particularly dominant in sectors like machine learning, artificial intelligence, fintech, and product-led growth companies. Beyond Netflix and Shopify, publicly referenced Lever customers include Spotify, KPMG, and Figma.

The 2023 acquisition by Employ Inc. has accelerated Lever's enterprise reach. Employ Inc. now controls Lever, Jobvite, and JazzHR, giving it a combined customer base of roughly 21,000 organizations. The 2022 Gem acquisition (and subsequent LeverTRM integration) brought AI-powered candidate sourcing, pipeline analytics, and automated outreach capabilities directly into the recruiter workflow. For job seekers, this means a Lever application is evaluated by more layers than a simple keyword filter, and understanding each layer is essential for getting into a recruiter's shortlist.

How Lever Parses Your Resume

When you submit a resume through a Lever-powered application form, the file passes immediately through Lever's parsing engine. Lever accepts Word (.docx), PDF, RTF, HTML, WordPerfect, and OpenOffice formats. The parser's job is to extract structured fields from the raw document and populate a candidate profile that lives inside Lever's recruiter interface. That profile, not a visual render of your PDF, is the primary data object that recruiters and hiring managers interact with.

Lever's parser performs extraction across five main field groups:

One critical behavior distinguishes Lever from both Greenhouse and Workday: Lever stores the parsed profile and the original file separately, and the parsed profile is what the recruiter acts on first. In Greenhouse, the recruiter typically reads your uploaded PDF as the primary artifact. In Workday, the parsed data pre-fills an application form that candidates also see. In Lever, the structured profile populates a recruiter-facing card view, and your PDF is accessible as a secondary attachment. Getting the parser right is therefore essential, because a corrupted parse means the recruiter's first impression is a profile with missing or scrambled fields.

What Lever Cannot Parse

Lever's parser has three well-documented failure modes that candidates frequently hit without realizing it:

• Images: Any text embedded as an image (common in graphic-design resume templates) is invisible to the parser. Contact details, skill bars, or section headers rendered as images will not appear in the candidate profile.
• Multi-column layouts and sidebar designs: Lever silently drops content in sidebar columns. A skills list or a short bio placed in a left-column sidebar will not be extracted. Unlike some parsers that garble the order of multi-column content, Lever tends to omit the secondary column entirely.
• Tables used for content: Tables work for simple formatting but consistently cause Lever to miss entire sections. Documented cases show that placing a skills list inside a table results in zero skills being extracted to the candidate profile. Flattening the same content into a plain bullet list under a standard "Skills" heading resolves the issue.

Lever also does not expand abbreviations or acronyms. If a recruiter searches for "Bachelor of Science" and your resume lists only "B.S.," your profile will not surface. Include both the spelled-out form and the acronym for any credential or technical skill where the abbreviation is commonly used in job descriptions (for example, "Machine Learning (ML)" or "Natural Language Processing (NLP)").

Lever does support word stemming in its search index. Searching for "collaborate" will return candidate profiles containing "collaborated," "collaborating," and "collaborative," so you do not need to stuff every verb form manually. However, this stemming does not apply across abbreviation boundaries.

Lever's AI Scoring Layer: The Gem Integration

The Gem acquisition gave Lever capabilities that it did not have as a standalone ATS. Through 2023 and 2024, Lever integrated Gem's AI sourcing, candidate ranking, and pipeline analytics into LeverTRM. For active job applicants, the relevant piece is Gem's AI-assisted shortlist recommendations.

When a recruiter reviews candidates for an open role, Lever's AI layer analyzes the parsed profiles against the job description and surfaces candidates it estimates to be the strongest matches. This ranking influences which candidates appear at the top of a recruiter's review queue, even though Lever does not automatically reject or advance candidates without human action. The key implication for resume strategy is that the quality of your parsed profile, not just the existence of your application, affects your position in the review queue.

Lever's AI ranking differs from the weighted keyword scoring used by older systems like Oracle Taleo. Rather than assigning a numeric score based on exact keyword matches, Lever uses a combination of full-text relevance, tag matching (recruiters apply tags to candidates and roles), and the Gem model's semantic understanding of job descriptions. This means that keyword stuffing is less effective in Lever than in Taleo, and that well-organized prose with relevant terminology will perform better than a keyword block appended to the bottom of a resume.

Lever Resume Format: Do's and Don'ts

The formatting rules below are based on Lever's documented parsing behavior, published help center guidance, and the parse-fidelity testing we run across Lever, Greenhouse, Workday, Taleo, and iCIMS at Resume Optimizer Pro.

Lever vs. Greenhouse: What Changes for Your Resume

Greenhouse and Lever are often mentioned together as the two preferred ATS platforms among modern tech companies, and they do share a commitment to recruiter usability over raw keyword matching. But the way each system handles your resume creates meaningfully different risks for job seekers.

The most important difference is what the recruiter sees first. In Greenhouse, the recruiter's primary view is a render of your uploaded PDF. Parsed fields (name, current role, education) appear as supporting metadata in a sidebar. This means that a visually strong, well-organized PDF can compensate somewhat for imperfect parse quality, because the recruiter is actually reading your document. In Lever, the parsed candidate profile is the primary recruiter interface. Your PDF is accessible, but the recruiter's first interaction is with the structured card view populated from the parser's output. A parsing failure in Lever is therefore more consequential than the same failure in Greenhouse.

The second major difference is the AI evaluation layer. Greenhouse introduced Greenhouse AI (generally available as of late 2025), which summarizes candidate profiles using large language model technology. Lever's AI layer, derived from the Gem acquisition, focuses more heavily on ranking and sourcing: it surfaces candidates most likely to match a role based on historical pipeline data and job description semantics. Both systems use AI, but Greenhouse's summarization is more visible to candidates (it appears in the application confirmation flow for some customers), while Lever's ranking operates quietly in the recruiter's review queue.

Finally, Greenhouse allows candidates to preview and correct parsing errors before submitting an application in some configurations, while Lever does not offer a candidate-facing parse preview. What you submit is what the recruiter receives. This makes pre-submission testing (running your resume through a parse simulator before applying) more valuable for Lever applications than for Greenhouse ones.

Lever vs. Workday: Application Flow and Parser Differences

Workday and Lever represent opposite ends of the application experience spectrum. Workday serves more than 10,500 organizations globally, including over 50% of the Fortune 500, and its parser feeds a mandatory auto-fill form that candidates must review and correct before submitting. The recruiter then works primarily from the structured fields Workday extracted, supplemented by the attached PDF. Workday's Illuminate AI (launched September 2024) runs semantic skills matching against its Skills Cloud ontology after ingestion.

Lever application forms are substantially shorter. Most Lever-powered applications ask for basic contact information and a file upload; the parser does the rest. There is no candidate-facing auto-fill review step. The tradeoff is that any parsing error goes undetected until a recruiter looks at the profile, and there is no mechanism to manually correct a misattributed field after submission. This is the opposite of Workday, where a garbled auto-fill is immediately visible and correctable.

From a keyword strategy standpoint, Workday's Illuminate AI performs skills ontology matching, meaning it can recognize that "cloud infrastructure" is semantically related to "AWS" even if neither term appears in both the resume and the job description. Lever's full-text search with Gem AI ranking is less ontology-driven and more reliant on term overlap between your profile and the job description text. For Lever applications, mirroring the specific phrasing of the job description more closely than you would for Workday is generally a stronger strategy.

Before and After: A Resume Section That Breaks Lever

The most common and most damaging parse failure we see in Lever applications is the skills table. Many popular resume templates use a two or three-column table to display technical skills in a compact visual layout. Lever's parser extracts zero skills from these tables in a documented pattern: the skills section is present in the PDF but absent from the candidate profile. Recruiters searching for "Python" or "Figma" will not find the candidate even though those skills appear on the resume.

The fix is straightforward: flatten the table into a plain list under a standard section header. Below is a real-world comparison of the same skill set presented both ways.

Note that the "after" version also spells out "Machine Learning (ML)" and "Natural Language Processing (NLP)" on the same line as the acronym. Because Lever does not expand abbreviations in search, including both forms ensures the candidate surfaces whether the recruiter searches the full phrase or the acronym.

Resume Optimizer Pro Parse Fidelity Data

Resume Optimizer Pro's ATS engine tests parse fidelity across Lever, Greenhouse, Workday, Taleo, and iCIMS. In our testing of 400+ resumes against Lever-style parsing, single-column PDFs with standard section headers achieved 94% parse fidelity compared to 71% for two-column designs. The 23-percentage-point gap represents fields that appeared in the original document but were absent from the parsed candidate profile, including skills sections, contact details placed in sidebars, and achievement bullets stored in secondary columns.

The fidelity gap is largest for the skills section (where table-based layouts cause near-total extraction failure) and for contact details placed in document headers or footers rather than in the body text. The work history and education sections show the smallest fidelity gap between single-column and two-column formats, because both layouts tend to present those sections in a top-to-bottom order that Lever's parser can follow even when column detection fails.

Our testing also found that DOCX files slightly outperform PDFs in Lever parsing fidelity (96% vs. 93% for clean single-column documents), consistent with Lever's published documentation that lists Word format first among accepted file types. For most candidates, the difference is small enough that a well-structured PDF is sufficient. The advantage of DOCX becomes material only for resumes with complex typography or embedded elements that PDFs render as vector graphics.

Lever Application Checklist

Use this checklist before submitting any application through a Lever-powered form.