Guidance
How Jetstream Jobs's AI works under the hood — the research-backed rules, bias counter-strategies, and coaching frameworks it follows when generating your materials.
Bias Awareness
Hiring bias is real and measurable. Jetstream Jobs actively watches for and corrects these patterns in every document it generates.
Gender Bias
Women's accomplishments are often attributed to luck or teamwork. Women face "prove it again" bias — needing to demonstrate competence repeatedly — and their leadership is perceived as "collaborative" rather than "visionary."
Research:
Williams & Dempsey, What Works for Women at Work (2014). Documented four patterns of gender bias in professional settings: Prove-It-Again, the Tightrope, the Maternal Wall, and Tug of War. Women must provide more evidence of competence to be judged as equally competent.
How Jetstream Jobs uses this:
- Enforces ownership language: "Built", "Architected", "Shipped" — never "Helped", "Contributed to", "Assisted"
- Flags passive voice and hedge words: "helped ensure", "was involved in", "participated in"
- Front-loads measurable accomplishments before soft-skill descriptions
- Quantifies everything — numbers are harder to argue with
Instead of:
"Helped ensure the platform was reliable"
Say:
"Achieved 99.99% uptime across production services"
Name and Racial Bias
Identical resumes receive different callback rates based solely on the name at the top. "Culture fit" evaluations often mask bias — interviewers mistake personal comfort from shared backgrounds as professional suitability.
Research:
Bertrand & Mullainathan, "Are Emily and Greg More Employable than Lakisha and Jamal?" (NBER, 2003). Sent 5,000 resumes to 1,300 job ads. White-sounding names received 50% more callbacks than identical resumes with Black-sounding names. The gap was equivalent to eight additional years of experience.
How Jetstream Jobs uses this:
- Leads with metrics-first framing — quantified impact is objective and harder to dismiss
- Coaches redirecting "fit" evaluations to professional values: ownership culture, delivery cadence, quality standards
- Structures answers around professional culture values, not social-similarity signals
Instead of:
"I like hiking and craft beer" (social-similarity fit)
Say:
"I care most about ownership culture and quality standards. At my last role that looked like a team shipping weekly with near-zero customer-facing incidents."
Socioeconomic Class Bias
Hiring managers detect socioeconomic class from seven words of speech and then rate higher-class speakers as more competent, assigning them higher starting salaries with zero information about qualifications.
Research:
Kraus et al., "Evidence for the Reproduction of Social Class in Brief Speech" (PNAS, 2019). Yale researchers found that hiring managers accurately assessed speakers' socioeconomic backgrounds from just seven words — and then rated higher-class speakers as more competent and offered them higher salaries, independent of actual qualifications.
How Jetstream Jobs uses this:
- Detects deferential framing (treating interviewer as evaluator vs. peer) and reframes as peer-mode interaction
- Flags hedge language: "kind of", "sort of", "I think maybe"
- Coaches opening lines — the first seven seconds have outsized impact
- Steers small talk toward domains where you have genuine expertise: side projects, conference talks, published work
- Leads with the builder narrative — "built the team and shipped the product" reads as founder-energy regardless of background
Instead of:
"I was given the opportunity to..." (deferential)
Say:
"I chose the approach, hired the team, and shipped v1 within nine months." (peer-mode, ownership)
"Culture Fit" Bias
"Culture fit" is the most common reason interviewers give for rejecting a candidate, yet research shows it frequently reflects shared socioeconomic background rather than professional compatibility.
Research:
Rivera, Pedigree: How Elite Students Get Elite Jobs (Princeton, 2015). In a two-year ethnography of elite hiring at law firms, banks, and consultancies, Rivera found that interviewers screened candidates on extracurricular activities signaling "intensive leisure" (sailing, not working through college), "polish" (a class-coded behavioral calibration), and "fit" (shared hobbies and backgrounds). When interviewers said they "clicked" with a candidate, they meant shared background — same sports, same grad school, same vacation spots. They misread personal comfort as professional suitability. "In many instances, culture fit is a form of discrimination."
How Jetstream Jobs uses this:
- Coaches redirecting "what do you do for fun?" from social-similarity territory to professional values
- Prepares answers that shift from social fit to values fit: ownership, shipping cadence, quality bar
- Frames "fit" answers around what kind of team culture you build, not personal lifestyle signals
When asked "what do you do for fun?"
"I like hiking" (social-similarity screening)
Redirect to professional values:
"I care most about ownership culture — teams where people ship and take pride in what they build. At my last company that looked like weekly releases and a team that genuinely cared about quality."
Industry and Employer Bias
Candidates from nonprofits, government, academia, or less-recognized employers face automatic assumptions that their work was small-scale, non-technical, or unserious — regardless of actual scope.
Research:
Rivera (2015) found interviewers used employer prestige as a shortcut for competence — candidates from recognized brands were assumed competent until proven otherwise, while candidates from lesser-known employers had to prove competence from scratch. This "resume whitening" effect extends beyond race to industry and employer type.
How Jetstream Jobs uses this:
- Scans materials for words that trigger industry bias ("nonprofit", "association", "government") and replaces with scale metrics
- Uses action-oriented language: "Shipped" not "delivered", "Own" not "responsible for"
- Names specific technologies and tools rather than vague categories
- Frames lesser-known employers as "built from zero" narratives — startup-within-enterprise is a respected pattern
Instead of:
"Manager at a nonprofit education organization"
Say:
"Scaled platforms to 75K monthly active users across 2,000+ organizations"
Age Bias
Graduation dates, early career details, and certain experience framing can trigger age-based assumptions that work against candidates in either direction — too junior or too senior.
Research:
Neumark et al., "Is It Harder for Older Workers to Find Jobs?" (NBER, 2015). Sent 40,000 resumes to 13,000 positions. Older applicants (ages 49-51 and 64-66) received significantly fewer callbacks, with the strongest bias against older women. Resume design choices — graduation dates, early roles, experience-duration framing — all contributed to age signaling.
How Jetstream Jobs uses this:
- Omits graduation dates from resumes
- Frames experience in terms of impact, not duration: "grew the team to 20" not "20 years of experience"
- Never uses the word "still" — it is defensive and implies surprise at continued relevance
- Focuses on current tools and recent accomplishments rather than career chronology
Instead of:
"With over 16 years of experience, I still stay current with modern frameworks..."
Say:
"Selected the tech stack for the latest platform rebuild; invited to guest lecture at a top-10 graduate program"
Resume Rules
The rules the AI follows when writing and reviewing resume bullets.
Ownership Language
Every bullet uses active, ownership-oriented verbs. The actor is always the subject.
| Never use | Use instead |
|---|---|
| Helped | Built |
| Contributed to | Shipped |
| Assisted | Led |
| Was involved in | Owned |
| Participated in | Designed |
| Responsible for | Grew / Scaled / Achieved |
| Delivered | Shipped |
Why this matters:
Research on gender bias in hiring (Williams & Dempsey, 2014) shows hedge words disproportionately harm women and underrepresented candidates. Ownership language is a concrete counter-strategy. Jetstream Jobs scans every bullet for passive and deferential verbs and suggests replacements.
Achiever vs Do-er Framing
Every bullet must position you as an achiever who changes the trajectory of the business — not a do-er who executes well. The test: ask "so what?" after every bullet. If it doesn't answer that, it needs a business outcome.
Do-er (just describes the work):
"Expanded the team, hiring additional staff."
Achiever (ties to business outcome):
"Expanded the team by 40%, reducing time-to-market for key features and increasing release frequency."
Do-er:
"Managed the product roadmap and coordinated with stakeholders."
Achiever:
"Owned the product roadmap, prioritizing three initiatives that drove 25% revenue growth and doubled user retention."
How Jetstream Jobs uses this:
When generating or reviewing bullets, the AI applies the "so what?" test to every line. If a bullet describes an action without tying it to a business outcome (revenue, retention, cost savings, time-to-market, strategic advantage), it rewrites to add the impact.
Bullet Structure
Rules for every bullet:
- One sentence per bullet — keep them scannable and atomic
- Lead with a strong action verb
- Include a metric, outcome, or specific context
- 1-2 lines; never exceed 3 lines
- No first-person pronouns — "Built platform" not "I built platform"
- Vary verbs across bullets — no repetition
Bullet formula options:
- Accomplished X as measured by Y by doing Z
- Action verb + accomplishment + outcome
Lead with Metrics
Numbers are harder to argue with and signal competence objectively. The first bullet of each job should be a high-level contribution with scale.
Bullet ordering strategy:
- Scale / metrics
- Technical accomplishments
- Product impact
- Leadership scope
- Process improvements
How Jetstream Jobs uses this:
When generating resumes, the AI reorders bullets within each role to lead with the highest-impact metric. It also checks that the first bullet for each position includes a concrete number — users, revenue, team size, or percentage improvement.
Interview Coaching
Frameworks the AI uses when preparing interview stories and coaching responses.
The Three-Layer Framework
Every interview story must have three layers: decision (yours), execution (your team's), system (yours).
The sentence pattern:
"I [decided/designed/set the direction]. [Name or 'my team'] [built/shipped/implemented]. The result was [outcome]."
Layer 1 — The Decision (say "I"):
"I identified the bottleneck and proposed the redesign." "I made the case to Finance for the $58K investment."
Layer 2 — The Execution (say "my team" or name the person):
"My senior developer built the core module in a 6-week sprint." "The design team shipped 12 iterations based on user feedback."
Layer 3 — The System (say "I"):
"I structured the work so my team could focus on the core product while I owned the integrations." "I built the feedback pipeline so customer issues automatically become prioritized work items."
How Jetstream Jobs uses this:
When preparing interview stories, the AI checks that each story contains all three layers. If a story only describes execution ("we built X"), it prompts you for the decision you made and the system you put in place. This framework prevents two common failure modes: hiding behind "we" (no credit) and claiming everything (no delegation).
"I" vs "We" Calibration
| Say "I" for | Say "my team" / names for |
|---|---|
| Decisions, direction-setting | Execution, shipping |
| Strategy, design choices | Implementation details |
| Hiring calls, budget decisions | Day-to-day delivery |
Key rule:
Never hide behind "we" when the interviewer wants to know what you did. That's as damaging as taking credit for everything.
Handling the "No Peers" Gap
When asked "who did you partner with on that decision?" and the honest answer is nobody:
"That was my call. I was the most senior leader in that area, so those decisions were mine to make. One of the things that attracts me to [Company] is having peers at that level to pressure-test those decisions with."
This is honest, doesn't invent fake collaboration, and turns the gap into a reason you want this job. Never use "level up" or "learn from smart people" framing — frame as better decisions through collaboration, not personal development.
First 7 Seconds Matter
Hiring managers detect socioeconomic class from seven words of speech and form bias before you say anything substantive. The first seven words have outsized impact.
Research:
Kraus et al. (PNAS, 2019). Pronunciation patterns in just seven words were enough for hiring managers to assess socioeconomic class — and then assign higher competence ratings and salaries to higher-class speakers, with no information about actual qualifications.
How Jetstream Jobs uses this:
- Coaches your "tell me about yourself" opening until it is polished and confident
- Treats the interviewer as an intellectual peer — asks questions back mid-answer, gently challenges premises
- Frames judgment calls as peer briefings, not defenses to a judge
- Includes "the thing I'd do differently" — signals intellectual openness and self-awareness
Words to Avoid
The AI flags and rewrites these words whenever they appear in your materials.
Banned Buzzwords
These words are filler that every candidate uses. They tell instead of showing.
The fix:
Instead of telling, show. Replace "skilled communicator" with "keynote speaker at three industry conferences." Replace "results-driven" with the actual result.
Bias Trigger Words
These words activate specific hiring biases. Jetstream Jobs scans all materials for them before submission.
These trigger the assumption that your work is small-scale or non-technical. Use the organization name alone or describe by scale metrics instead.
Defensive and Filler Words
| Word | Why it's flagged |
|---|---|
| still | Defensive — implies surprise at continued relevance |
| actually | Filler that weakens statements |
| really | Qualifier — delete it, the sentence is stronger without it |
| basically | Qualifier — sounds unsure |
| very | Qualifier — find a stronger word |
| somewhat | Qualifier — be definitive |
| kind of | Hedge — signals low confidence |
| sort of | Hedge — signals low confidence |
Passive Voice Patterns
Passive voice hides the actor and weakens impact. The AI rewrites all passive constructions.
| Passive | Active |
|---|---|
| "Was given oversight of..." | "Took ownership of..." |
| "The platform was built..." | "Built the platform..." |
| "Improvements were made..." | "Shipped improvements that..." |