The Cognitive
Your hiring desk, run by Claude.
Sourcing first. What to type, and what comes back. One role, from the brief to the ATS.
Connect Claude
- Start your free 7-day trial at The Cognitive.
- In Claude, open Settings, then Connectors, and add a custom connector:
https://app.thecognitive.io/mcp - Sign in and type any line below. Claude asks before every email, reveal, stage move and rejection.
Sourcing
05 to 16
05Plain-language searchWrite the brief the way you would tell a colleague.
Senior data engineers, Austin area. Must have run Spark in production on payments or billing data. 5+ years. Nice to have: dbt. Skip anyone at a staffing agency or consultancy.
A brief has 5 parts
- Role and level. Senior data engineer.
- Place. The city, plus "area" or "within commuting distance". Or "remote in the US".
- 2 or 3 must-haves. Written as things they did: "has run Spark in production", not "Spark".
- Nice to have. What tips a tie.
- Who to skip. Staffing agencies, your clients, people you already passed on.
Weak vs strong
Data engineer Austin Spark Airflow 5 years
Senior data engineers, Austin area. Must have run Spark in production on payments or billing data. 5+ years. Skip staffing agencies.
No contact details yet. Next page when you ask.
More briefs to copy
Mid-market account executives in New York who have closed $50K+ deals on their own at a B2B SaaS company. 3+ years carrying a quota. Nice to have: sold to finance teams. Skip staffing agencies.
Customer success managers in the US, remote, who have owned renewals on a book of $1M+ at a SaaS company. Nice to have: onboarding. Skip support-only roles.
In-house technical recruiters in Toronto or Waterloo who have hired engineers at a company under 500 people. 3+ years. Skip agency recruiters.
Senior product designers in London who have shipped a design system used by more than one team. Nice to have: B2B SaaS. Skip agencies and freelancers.
06One clarifying questionA vague brief gets a question back, not a bad list.
Find me a marketing lead.
Pick an option and it searches. Or write the full brief and there is no question.
Before I search: which kind of marketing lead?
No credits used on a question.
07Scout from the JDSearch straight from a role you already set up.
Find new people for the Senior Data Engineer role.
Senior data engineer in Austin, TX, hybrid. Must have 5+ years, production Spark and Airflow. Nice to have dbt and Kafka.
People found and saved to the role. Want to change the brief and search again?
08Reopen a past searchGet last week's list back without paying again.
Reopen yesterday's Austin data engineer search.
Same people, ranked as before. 0 credits.
09Role poolEveryone you already found for a role, in one place.
Show me everyone we have already found for the Senior Data Engineer role.
| Name | Role |
|---|---|
| Dev Nair | Senior Data Engineer, fintech |
| Aisha Bello | Data Engineer II |
| Lena Wu | Data Platform Engineer |
Saved from past searches for this role. People you passed on are left out. Free.
10Already in your ATSKnow before you source someone you already have.
Which of these people are already in our ATS?
- Aisha Bello
- Lena Wu
Each one shows the stage and job they hold in your ATS.
11Shortlist and passPass once. They do not come back for that role.
Shortlist Dev Nair and Aisha Bello. Pass on Tom Becker.
Pass with a reason you would say out loud
- At a staffing agency now
- No Spark in production
- Too senior for the range
- Left their last two jobs inside a year
- Shortlisted: Dev Nair, Aisha Bello
- Passed: Tom Becker. Not shown again for this role.
12Shared tagsOne set of tags for the whole team.
Tag Dev Nair and Aisha Bello as "phone screen".
A tag set that stays useful
Tagged phone screen on both. The tag is now in the team library.
13Contact revealEmail when you need it. Never paid twice.
Get Dev Nair's email address.
- Claude tells you the cost and waits for your yes.
- Nothing is charged when nothing is found.
- Someone already revealed, here or in the app, costs nothing again.
Revealing an email costs 4 credits. Nothing is charged if none is found. Go ahead?
Yes.
Found and saved to your revealed list.
14Send from your mailboxThree lines from your own address.
Start email outreach to Dev Nair for the Senior Data Engineer role. Open with his ledger work.
The first message, 3 lines
- One fact from their work. Something they built or ran.
- The role. Title, place, pay range.
- A small ask. 20 minutes, a day.
Line 1, other roles
- You took the Austin region from zero to $3M.
- You hired the first 50 people at a seed-stage company.
- You owned renewals on a $2M book.
- You shipped the design system three product teams use.
Hi Dev,
You rebuilt a payments ledger on Spark.
We are hiring a Senior Data Engineer to own our billing pipelines. Austin, hybrid, $150K to $175K.
Open to 20 minutes next week?
This emails 1 person from your mailbox. Follow-ups stop when he replies. Send?
15Replies into the roleA yes lands on the role, not in an inbox.
Did anyone reply with interest on the Senior Data Engineer outreach?
Dev Nair replied: "Happy to talk, Thursday works."
Marked interested and added to Senior Data Engineer.
16Outreach statusWho got a message, who answered.
Show me the outreach status for the Senior Data Engineer role.
| Person | Status |
|---|---|
| Dev Nair | Interested |
| Priya Raman | Sent |
| Marco Silva | Scheduled |
Set up the role
01 to 04
01Role setupRole, job description and interview criteria in one go.
Set up a role for a Senior Data Engineer in Austin, hybrid, $150K to $175K. Must have production Spark and Airflow. Build the interview too.
What you getThe role, its job description and weighted interview criteria. You are the hiring manager unless you name someone.
02Job descriptionRewrite last year's JD.
Rewrite the job description for the Senior Data Engineer role. Lead with the problem they will own.
What you getA draft to read. Saved to the role when you say yes.
03Interview questionsSame questions for every interviewer.
Draft 5 interview questions for the Senior Data Engineer role, each with what a strong answer sounds like.
What you getQuestions built from the role's job description, each with what to listen for.
04Scoring criteriaAgree what good looks like before round one.
Propose scoring criteria for the Senior Data Engineer role, weighted, with what a 5 looks like. Save them when I say yes.
What you getWeighted criteria, each scored 1 to 5, rolled into one score out of 100.
| Criterion | Weight | A 5 looks like |
|---|---|---|
| Production Spark | 30% | Tunes jobs, explains how they fixed skew |
| Pipeline ownership | 25% | Owned failures end to end, on-call included |
| Orchestration | 20% | Designs DAGs with retries and backfills |
| Data modelling | 15% | Models with analysts |
| Communication | 10% | Explains trade-offs plainly |
Interview
17 to 20
17Interview inviteThey book their own slot.
Invite Dev Nair to interview for the Senior Data Engineer role.
What you getClaude asks first. Then an email to book a slot. If the email cannot be sent, you get the booking link.
18Bulk inviteUp to 25 at a time.
Invite everyone on the Senior Data Engineer role with a resume score above 70.
What you getEveryone listed for your yes first. Refused up front if credits are short.
19Cancel and refundA cancelled slot costs nothing.
Cancel Thursday's interview with Omar Haddad.
What you getClaude asks first. He is told, and the credits held for it come back.
20Resume scoreEvery resume read against the job description.
Score every resume on the Senior Data Engineer role against the job description.
What you getA score and the reason for each. A resume that could not be read is marked unscored, not weak.
Decide
21 to 24
21Interview feedbackScores, gaps and the transcript.
Show me Maria Lopez's interview feedback and why she scored what she did.
What you getA score per criterion, the overall score, strengths, gaps and the full transcript.
22Score checkNo criteria behind a score? It says so.
Is Sam Patel's score reliable?
What you getIf the role had no criteria or the scoring failed, Claude says so instead of calling it a result.
23Move to the next roundOnly after the evidence.
Move Maria Lopez to the final round for the Senior Data Engineer role.
What you getClaude asks first. Then she moves, and it shows on her stage history.
24Reject with a reasonAlways written down.
Reject Sam Patel for the Senior Data Engineer role. Reason: no production Spark.
What you getNeeds a reason and your yes. The reason goes on his record.
ATS
25 to 28
25ATS importNo copying between tabs.
Import new applicants from our ATS now.
What you getClaude asks first. New people in each mapped job's invite stage come in. Nobody duplicated.
26ATS write-backStages move when you decide.
Write our decisions back to the ATS for the Senior Data Engineer role.
What you getClaude asks first. A note with the score and report, and the stage move your mapping sets, where your ATS allows it.
27Send to an ATS jobSourced people reach the ATS.
Send the people I shortlisted and revealed to the Senior Data Engineer job in our ATS.
What you getClaude asks first. Added as candidates. Someone already there gets the application, not a duplicate.
28Spreadsheet auto-syncNew rows every 15 minutes.
Turn on auto-sync for our applicants sheet.
What you getClaude asks first. Rows with a blank or "pending" Status come in every 15 minutes, and the Status is written back.
The desk
29 to 32
29Pipeline summaryWhere things stand, in one answer.
Where do things stand across all our roles?
What you getCandidates by stage, interviews by status, resume scores and interviews left.
30Stage historyWho moved, when.
Show me Maria Lopez's stage history.
What you getEach interview stage with its status and date.
31Re-score after a JD editResumes scored again on the new JD.
We added Kafka as a must-have to the Senior Data Engineer job description. Re-score the resumes against it.
What you getEach resume scored again against the new job description.
32Candidate data requestExport or delete on request.
A candidate at hana.kim@example.com asked what we hold on her. Export it.
What you getCounts of records, interviews, files and consent records. Admins only. Deleting needs her email typed again and your yes.