Pipeline01 uses AI to assist information preparation and role search. Gomes Godin LLC operates the service. This notice explains which tasks use a model, which use application rules, and the limits of both. It supplements the Privacy Policy, Candidate Participation and Consent Terms, and Acceptable Use and Fair Hiring Policy.
Optional resume autofill
You can complete candidate intake manually. If you choose PDF autofill and separately consent, Pipeline01 temporarily reads bounded text from the PDF, reduces direct contact details and candidate-name headers, and asks OpenAI to suggest structured work-history and professional fields. Company and education-provider names can remain in this input to support accurate form suggestions. Redaction may miss identifying material in free text.
Suggestions are not submitted automatically. You review, correct, remove, and supplement them before submitting. The new autofill flow does not retain the original PDF as an application attachment. It does not infer your desired compensation, job-search intent, or availability from a resume. A suggestion is not independent verification of a fact.
Candidate profile drafting
After you submit the form and confirm enrollment, a separate workflow uses saved application fields to prepare an employer-facing draft. It reduces direct identifiers and organization names before calling the model. The model is instructed to use supplied facts, omit identifying material, and leave missing facts unspecified. It does not receive a raw PDF as the source for this form-derived draft.
Pipeline01 staff review the draft before publication. The application requires active consent and a reviewed draft for publication. Opportunity preferences are taken from saved form fields rather than invented by the model. Claims about experience, outcomes, qualifications, and availability remain candidate-supplied unless a specific statement explains independent verification. Administrator review is not a background check.
Employer Fit Search
Fit Search has two stages:
- OpenAI extracts a structured brief from the employer's submitted role description: target role, required/preferred skills, domain, sales motion, market segment, experience target, location, work preferences, and a short summary. The model is instructed not to infer protected traits.
- Pipeline01 code compares that brief with eligible, published, reviewed profile fields and calculates results. Candidate profiles and private candidate contact details are not included in the model request used to extract the hiring brief.
The employer's role description is sent to the AI provider. Do not include candidate resumes, applicant names, health information, protected characteristics, trade secrets, or confidential client material without authority to disclose it. The application is designed not to save the submitted search description to its database, content logs, or analytics; provider handling and ordinary request/security metadata are separate.
Current ranking factors
The current implementation gives these components the following shares of its weighted calculation:
| Component | Weight | What it measures |
|---|---|---|
| Role alignment | 25% | Direct role/title overlap or adjacent role family |
| Required skills | 23% | Supported terms found in reviewed profile text |
| Domain or industry | 15% | Stated domain terms found in the profile |
| Experience target | 10% | Stated experience relative to a flexible years target |
| Sales motion | 10% | Relevant sales-motion terms found in the profile |
| Preferred skills | 7% | Preferred terms found in the profile |
| Market segment | 5% | Relevant market-segment terms found in the profile |
| Location and work preferences | 3% | Location/work terms matched against location, work-setting, and availability text |
| Presence of outcome/context evidence | 2% | Whether the reviewed profile contains a submitted win or other summary context; not independent verification of its quality |
The system normalizes selected sales-title synonyms and terms. A requested years target is not an automatic exclusion. When the employer does not specify a factor, the implementation can use a default contribution; the score is not simply a count of all known facts. Missing profile evidence can reduce a match even if the person has the skill in real life.
The weighted score is bounded and converted to a rounded rating out of 20. The labels are Strong fit (17–20), Good fit (14–16), Partial fit (10–13), and Weak fit (below 10). The underlying numeric score orders results; candidate references break exact ties. The service currently considers at most 200 returned eligible profiles and returns at most five results. Database retrieval limits mean the search is not a promise to evaluate every person in the network.
Neither the rating nor its label is a probability of hiring success, an independently validated employment assessment, a quality certification, or a determination that a candidate meets every requirement. A missing result is not a rejection or a finding of unfitness. Short explanations summarize the available matches and gaps; they do not include all candidate experience or every part of the scoring calculation.
Human review and lawful hiring
Pipeline01 does not use the score to make an offer, reject an application, determine compensation, or make a final employment decision. Employers must review relevant evidence, assess the accuracy and job relevance of extracted requirements, consider reasonable accommodations, and make their own lawful decisions. Do not use a score as the sole basis for excluding someone or treat missing evidence as proof of inability.
AI extraction, published profile wording, and keyword matching can be wrong, incomplete, inconsistent, or sensitive to phrasing. Experience, geography, education-related text, and language can act as proxies for protected characteristics. Removing direct identifiers does not establish fairness or eliminate discrimination risk.
An assistive description or human involvement does not by itself exempt a tool from automated-employment-decision, privacy-profiling, notice, bias-audit, or employment laws. This notice is not a jurisdiction-specific pre-use notice or a substitute for an audit required by law. Employers and Pipeline01 must assess obligations applicable to their respective activities before covered use.
Provider handling
These workflows use OpenAI API requests with response storage disabled. This setting is not Zero Data Retention. OpenAI states that API content is not used for training by default unless the customer opts in; default abuse-monitoring retention can include content for up to 30 days, with exceptions. Pipeline01 has not established approved Zero Data Retention through this notice. See OpenAI data controls. Account settings and provider terms must match any published commitment.
Choices and concerns
Candidates can decline PDF autofill and complete the form manually. The current listing process requires separate consent to AI-assisted profile drafting and participation. It does not provide a separate automated option to stay listed while excluding your profile from Fit Search. You can contact us to request correction, discuss an explanation or human review, or withdraw from the network. We do not promise an AI-free listing alternative that the service does not currently support.
Employers may use the ordinary board without submitting a role description to Fit Search. Report inaccurate requirements, unsafe output, or suspected discriminatory behavior to tgodin@gomesgodin.com. Do not include private applicant records in a report unless we arrange an appropriate secure method.
If the input categories, ranking factors, model role, candidate visibility, or decision use changes materially, Pipeline01 must reassess this notice and the permissions and controls required for the change.