BTS - how graph helps identify the missing context

A lot of you must be wondering how does @graph actually audit a Data & AI resume.

So thought of sharing what's under the hood.

Every bullet point gets checked against 7 things:

1. What: the work itself. What was built, changed, fixed or delivered.

2. How: the tools, languages, platforms or methods used.

3. Scale: how big the work was. Data volume, record counts, number of pipelines or users, budget. "Large-scale" with no figure is vague. And a percentage never counts as scale.

4. How much: a measured result. A percentage, time saved, cost, a before and after. "Improved" or "faster" with no number is vague.

5. Why: the problem or purpose behind the work. What was wrong before.

6. Your role: was it actually your work? Built, led, owned pass. Helped, supported, "worked on" are vague.

7. Outcome: who it helped. Users, the team, the business, a system. If the effect is only implied, it's vague.

Now, let's take a very common bullet from a Data Engineering resume:

"Helped build large-scale data pipelines that improved reporting speed."

and run it through @graph's core eval criteria:

• What: data pipelines ✅

• How: missing ❌

• Scale: "large-scale" ⚠️ vague

• How much: "improved" ⚠️ vague

• Why: missing ❌

• Your role: "helped" ⚠️ vague

• Outcome: implied ⚠️ vague

1 out of 7. And tbh, you'll be surprised how common this is across roles / levels.

And now imagine @graph probing you and helping plug all the missing context - then the bullet would look something like:

"Built 12 Airflow and dbt pipelines processing 40M records a day, replacing a manual Excel process that took finance 3 days every month. Cut month-end reporting from 3 days to 4 hours, so the team could close the books a week earlier."

My bet: the person usually knows all of this already.

The tools, the numbers, the reason it mattered. It's just never articulated well enough. That's why graph first asks questions instead of simply polishing what already exists.

A few patterns I keep noticing from early users:

→ "Why" is the most commonly missing piece. People describe the solution and skip the problem.

→ Percentages get mistaken for scale. "Improved accuracy by 15%" tells me the result, not how big the work was.

→ "Helped" and "supported" quietly undersell people who actually led the work.

Try for free it on your own resume.

Let me know in comments, what was your resume missing?