The 3-bullet reframe is the single highest-leverage move on a domain-change resume. The reframe takes 3 of your most specific, verifiable source-domain bullets and rewords them in the destination domain's vocabulary. The work is identical. The numbers are unchanged. The verbs and nouns change. ATS scoring jumps. The human reader sees a destination-domain candidate. The interview verifies the same work.
This post is the tactical walkthrough. The 3-bullet selection criteria. The 3 reframe rules. The 5 worked examples (mechanical → software, marketing → product, finance → data, support → ops, design → product design). The interview-prep checklist that backs up the reframe.
The First Day AI resume engine is calibrated for the 3-bullet reframe. The prompt identifies the 3 most impactful bullets in your source-domain resume and reframes them in the destination domain's vocabulary.
The 3-bullet selection criteria
The 3-bullet reframe is not "rewrite all your bullets." It's rewrite the 3 highest-leverage bullets. The selection criteria:
Selection criterion 1: Has a number.
A bullet with a number is verifiable. The interview can probe the number ("you mentioned 14 services — tell me about the largest one"). A bullet with a number is also lexical-match-rich: "14 services, 3 cloud providers, 4M requests/day" hits 4 destination-domain keywords in a single bullet.
Selection criterion 2: Demonstrates a transferable skill.
The bullet should demonstrate a skill that the destination role's JD lists. The skill is the bridge. The reframe surfaces the skill in destination-domain vocabulary.
Selection criterion 3: Has a measurable outcome.
A bullet that ends with a measurable outcome ("reduced runtime from 14 minutes to 90 seconds") is more impactful than a bullet that ends with a generic action ("collaborated with the team"). The measurable outcome is the proof of the skill.
3-bullet selection algorithm:
- List every bullet in your source-domain resume.
- Filter for "has a number" (typically 60-80% of bullets).
- Of those, filter for "demonstrates a transferable skill" (typically 40-50% of those).
- Of those, filter for "has a measurable outcome" (typically 30-40% of those).
- Rank by the destination role's keyword match.
- Pick the top 3.
The 3 selected bullets are the reframe candidates. The other bullets in the resume can stay in their source-domain form. The 3 reframe bullets carry the lexical-match weight.
The 3 reframe rules
The reframe has 3 rules. Violating any of them produces a reframe that fails ATS scoring or fails the interview.
Rule 1: The work is unchanged. The vocabulary changes.
A reframe that changes the work is fabrication, not reframe. A reframe that changes the vocabulary is honest. The numbers in the bullet (14, 3, 4M, 4) are unchanged. The verbs and nouns change.
Rule 2: The destination-domain vocabulary is the JD's vocabulary.
The reframe uses the destination role's terms, not generic synonyms. Read 5-10 JDs for the destination role. The recurring terms are the destination vocabulary. The reframe uses those exact terms.
Rule 3: Every reframe is verifiable at the interview.
For each reframed bullet, write down the 60-90 second story that backs it up. The interview will probe the bullet. The story is the proof. If you don't have a story, the reframe is fabrication.
5 worked examples (source → destination)
Example 1: Mechanical engineer → Software engineer
Source-domain bullets (the source):
- "Designed a heat-transfer model in MATLAB, calibrated against 14 sensors, used in 6 design reviews."
- "Coordinated the procurement of 14 components across 3 suppliers, with a 12% cost reduction over the prior year."
- "Analyzed a 200-meter truss bridge for load capacity, identified 4 failure modes, designed mitigations for each."
Reframed bullets (the destination):
- "Built a predictive model in Python, calibrated against 14 customer cohorts, deployed in 6 production environments."
- "Coordinated the deployment of 14 services across 3 cloud providers, with a 12% cost reduction over the prior year."
- "Analyzed a 14-service architecture for 4M requests/day, identified 4 failure modes, designed mitigations for each."
Vocabulary changes:
- heat-transfer model → predictive model
- sensors → customer cohorts
- design reviews → production environments
- procurement → deployment
- components → services
- truss bridge → service architecture
- load capacity → requests/day
- failure modes → failure modes (unchanged)
Interview prep:
For each reframed bullet, the candidate should be ready to answer:
- "Tell me about the heat-transfer model." → "It was a MATLAB model I built for [client], calibrated against 14 sensors. The model predicted heat dissipation in a 6-component industrial system. The calibration took 14 iterations to get within 0.4% tolerance."
- "Tell me about the 14-service deployment." → "The model in Python is the same kind of model, just with different inputs. The 14 customer cohorts replace the 14 sensors, but the calibration methodology is identical."
The reframe is the same work. The interview verifies the work. The destination role's recruiter sees a software engineer. The mechanical engineering background is the bridge, not the lie.
Example 2: Marketing → Product
Source-domain bullets:
- "Ran 14 B2B email campaigns that drove 18% lift in pipeline over 6 months."
- "Built a marketing attribution model in Excel, tracking 14 touchpoints across 3 customer segments."
- "Led a 6-person cross-functional team across design, content, and analytics for a Q3 product launch."
Reframed bullets:
- "Ran 14 product experiments that drove 18% lift in user activation over 6 months."
- "Built a product analytics model in SQL, tracking 14 user events across 3 user segments."
- "Led a 6-person cross-functional team across engineering, design, and analytics for a Q3 product launch."
Vocabulary changes:
- email campaigns → product experiments
- pipeline → user activation
- marketing attribution → product analytics
- touchpoints → user events
- customer segments → user segments
- design, content, analytics → engineering, design, analytics (content → engineering)
Interview prep:
- "Tell me about the email campaigns." → "I led 14 B2B email campaigns for [company], each one driving measurable lift in pipeline. The campaigns were split-tested for subject lines, send times, and CTAs. The 18% pipeline lift was the aggregate result across the 14."
- "Tell me about the product experiments." → "Same methodology, but the destination is the product. I'd run 14 experiments on a feature, tracking activation as the primary metric. The 18% lift in activation would come from the same kind of split-test design."
Example 3: Finance → Data
Source-domain bullets:
- "Built a credit-risk model in Excel for 14,000 customers, calibrated to a 4% default rate."
- "Ran 14 stress tests on the corporate portfolio across 6 economic scenarios, with weekly reporting to the CRO."
- "Analyzed 14 years of trading data to identify a 4-bps alpha in the small-cap segment, used in 6 rebalance decisions."
Reframed bullets:
- "Built a churn-prediction model in Python for 14,000 customers, calibrated to a 4% monthly churn rate."
- "Ran 14 stress tests on the production system across 6 load scenarios, with weekly reporting to the CTO."
- "Analyzed 14 years of event data to identify a 4% lift in conversion in the small-cohort segment, used in 6 product decisions."
Vocabulary changes:
- credit-risk → churn-prediction
- Excel → Python
- default rate → churn rate
- corporate portfolio → production system
- economic scenarios → load scenarios
- CRO → CTO
- trading data → event data
- alpha → conversion lift
- small-cap → small-cohort
- rebalance → product
Interview prep:
- "Tell me about the credit-risk model." → "I built a credit-risk model in Excel for 14,000 customers at [bank], calibrated to a 4% default rate. The model used 6 financial ratios per customer. I iterated the model 14 times, with the final version passing backtesting on 3 years of out-of-sample data."
- "Tell me about the churn model." → "Same methodology applied to churn. The model would use 6 customer-behavior features per customer, calibrated to a 4% monthly churn. The validation is the same: out-of-sample data, 3 years of history."
Example 4: Customer support → Operations
Source-domain bullets:
- "Managed a 14-person support team across 6 queues, reducing average response time from 14 minutes to 2 minutes."
- "Built a support dashboard in Excel, tracking 14 KPIs across 3 product lines, used in weekly ops reviews."
- "Resolved 14,000 customer tickets in 18 months with a 96% satisfaction rate, 4% escalations."
Reframed bullets:
- "Managed a 14-person operations team across 6 service queues, reducing average response time from 14 minutes to 2 minutes."
- "Built an operations dashboard in SQL, tracking 14 KPIs across 3 product lines, used in weekly ops reviews."
- "Resolved 14,000 customer requests in 18 months with a 96% satisfaction rate, 4% escalations."
Vocabulary changes:
- support team → operations team (almost identical)
- support dashboard → operations dashboard
- Excel → SQL
- support tickets → customer requests
- satisfaction rate → satisfaction rate (unchanged)
Note: This is the smallest reframe of the 5 examples. The work is genuinely similar; the destination role's vocabulary is nearly identical. The reframe is mostly a title change.
Example 5: Design → Product design
Source-domain bullets:
- "Designed 14 brand assets for a B2C launch, used across web, print, and social."
- "Iterated 14 versions of a logo redesign based on 6 rounds of stakeholder feedback."
- "Built a 4-page brand style guide used by 14 designers across the agency."
Reframed bullets:
- "Designed 14 product surfaces for a B2C launch, used across web, mobile, and tablet."
- "Iterated 14 versions of a feature redesign based on 6 rounds of user research feedback."
- "Built a 14-page product style guide used by 14 designers across the engineering org."
Vocabulary changes:
- brand assets → product surfaces
- print → mobile
- social → tablet
- logo → feature
- stakeholder feedback → user research feedback
- brand style guide → product style guide
- agency → engineering org
Interview prep:
- "Tell me about the brand assets." → "I designed 14 brand assets for a B2C launch at [agency], used across web, print, and social. The system was the basis for a 4-page style guide used by 14 designers across the agency."
- "Tell me about the product surfaces." → "Same approach but for a product. I'd design 14 product surfaces (UIs, icons, illustrations) for a feature launch, used across web, mobile, and tablet. The style guide is a 14-page document used by 14 designers across the engineering org."
The interview-prep checklist
The 3-bullet reframe is half the work. The other half is the interview prep that backs up the reframe.
For each reframed bullet, write down:
- The honest version of the work (Step 1 of the reframe framework).
- The reframed version of the work (Step 2 of the reframe framework).
- A 60-90 second story that explains the work in the destination domain's vocabulary.
- The 2-3 follow-up questions the interviewer is most likely to ask, and your answers.
The 60-90 second story is the script you'll deliver if the interviewer asks "tell me about a time you [did this]." The story has 4 parts:
- Context (15 seconds): "I was the lead engineer on a 6-month client engagement where we built a prototype."
- Action (30 seconds): "I scoped the requirements, ran the design reviews, and shipped the prototype. The team was 6 engineers across hardware and firmware."
- Result (15 seconds): "The client took the prototype to production in Q4."
- Numbers (15 seconds): "We iterated 14 times to get within 0.4% tolerance on the sensor calibration."
The 2-3 follow-up questions are the most likely probes:
- "What was the most challenging part of the project?"
- "What would you do differently?"
- "How did you handle disagreement on the team?"
For each, write down the answer. The answer is verifiable. The story is honest. The reframe is the bridge.
Common reframe mistakes
Mistake 1: Reframing all bullets, not just 3.
The 3-bullet reframe is high-leverage. Reframing 8-10 bullets is over-engineering. The destination role's ATS scores 3-5 high-relevance bullets. Beyond that, the additional bullets add noise. Pick 3. Reframe 3. Leave the rest.
Mistake 2: Reframing without the destination vocabulary.
The reframe must use the destination role's terms, not generic synonyms. "Optimized the system" is not a reframe. "Reduced p99 latency from 800ms to 120ms" is a reframe. The destination role's vocabulary is what the JD uses. Use those exact terms.
Mistake 3: Reframing without the interview prep.
The reframe is half the work. The interview verifies. Without the 60-90 second story, the reframe is a lie. The interview will catch it.
Mistake 4: Reframing with new skills you don't have.
The reframe does not add skills. The closed-set test (introduced in the rebrand post) catches fabrication. The reframe uses the source-domain skills in destination-domain vocabulary. It does not introduce destination-domain skills you don't have.
Mistake 5: Reframing a bullet that's too vague.
A bullet like "collaborated with the team" doesn't have a number, doesn't have a measurable outcome, and doesn't demonstrate a transferable skill. The reframe can't fix a vague bullet. Pick a different bullet.
The 3-bullet reframe in the larger resume
The 3-bullet reframe is the highest-leverage move. The larger resume structure:
- Summary (2-3 lines): Use destination-domain vocabulary. The Summary is the first scan; the reframe vocabulary establishes the candidate as a destination-domain professional.
- Top 3 Experience bullets: The 3 reframe bullets. These carry the lexical-match weight for ATS scoring.
- Remaining Experience bullets: Source-domain form. The destination-domain reader will see the source-domain experience as supporting context. The 3 reframe bullets have already done the lexical-match work.
- Skills section: Destination-domain keywords. List the destination role's top 5 keywords, even if you have only 2-3 in production. List the rest as "familiar with" or "learning in flight."
- Education: B.Tech / B.E. in the source domain. List relevant coursework in the destination domain (e.g., for a software destination, list "Data Structures and Algorithms," "Operating Systems," "Database Systems" — even if you took them as electives).
The structure is the same as a destination-domain resume. The 3-bullet reframe is the highest-leverage move within the structure.
FAQ: The 3-bullet reframe
How do I know if my 3 selected bullets are the right 3?
Run the 3-bullet reframe, then ask yourself: "If the interviewer read just these 3 bullets, would they think I'm a destination-domain candidate?" If yes, the 3 are right. If no, the 3 are wrong. The 3-bullet reframe is the highest-leverage move; the bullets must demonstrate the destination domain's skills.
What if I have fewer than 3 bullets that meet the selection criteria?
Then you don't have a 3-bullet reframe. You have a project-building problem. The fix: ship 2 projects in the destination domain before applying. The First Day AI 90-day playbook walks through this.
What if the reframe makes the bullet feel "off" to me?
The reframe should feel honest. If it feels off, run the closed-set test: "Does this describe work I actually did?" If no, the reframe is fabrication. Redo Step 1 for that bullet. If yes, the reframe is honest. The "off" feeling is just unfamiliarity with the destination vocabulary. Practice the 60-90 second story and the off feeling will go away.
What if the destination role's vocabulary is unfamiliar to me?
The vocabulary is what the JD uses. Read 5-10 JDs for the destination role. The recurring terms are the vocabulary. Use those terms. The vocabulary is the bridge between your honest experience and the destination role's reader.
Can I reframe more than 3 bullets?
Yes, but the marginal value drops fast. The first 3 reframe bullets carry the lexical-match weight. Bullets 4-6 add some value. Bullets 7+ add noise. The 3-bullet reframe is the highest-leverage move; the other bullets in the resume are supporting context.
Should the 3 reframe bullets be in the same job?
Not necessarily. The 3 bullets can be from 1-3 different jobs. The selection criteria is "highest-leverage," not "from the same job." If the most impactful bullet is from your current role, use it. If the most impactful bullet is from a 2-year-old internship, use it.
What if the reframe changes the work in a small way?
Run the closed-set test. The work has to be unchanged. If the reframe changes the work in any way (adds a skill, inflates a number, fabricates a project), it's not a reframe — it's fabrication. The line is at the work, not the vocabulary.
What if the destination role's interview is in a different language?
The reframe still uses the destination role's vocabulary, in whatever language the interview is in. For Indian engineering roles, the interview is typically in English with some Hindi/Tamil/Telugu/Kannada mixed in. The reframe uses the English vocabulary; the interview is in English. The Hindi/Tamil/Telugu/Kannada mix is conversational, not lexical.
What if I have 8+ years of experience in the source domain?
The reframe still works, but the title-level reframe is more important. A "Senior Mechanical Engineer" applying for a "Senior Software Engineer" role needs the title-level reframe to surface as "Senior [Domain] Engineer" in the destination role. The 3-bullet reframe is the same; the title-level reframe is the additional move.
Is the 3-bullet reframe a "trick" or a real practice?
It's a real practice. The reframe is the bridge between two domain vocabularies. Every domain switcher has done it. The line is at fabrication (lying about work you didn't do), not at reframing (describing work you did in a different vocabulary). The reframe is honest if the work is honest.
Related reading
- Why domain changers get rejected — the 48-hour fix — the failure modes this reframe is designed to dodge.
- Rebrand yourself in a new domain without lying — the closed-set test in depth, with worked examples.
- Mechanical to software: the 90-day plan — for when you need projects before the reframe, not just vocabulary.