AI Engineering5 min
Tracking applications reliably
The hard part is not storing applications. It is that most of them never produce a clear ending.
Read moreCareer guides thatrespect your time
Updated
Page 9 of 11 · 205 guides in total
AI Engineering5 min
Scoring candidates responsibly
This is the one system in the stack where getting it wrong is a legal matter, not just a product one.
Read moreAI Engineering5 min
An assistant people come back to
Job searching is stressful, often for months. That fact should shape the product more than any model capability does.
Read moreAI Engineering4 min
Analytics on a very small sample
Forty applications is not a dataset. Most job search analytics present noise with a confident chart around it.
Read moreAI Engineering5 min
Parsing a job description
Everything downstream — matching, scoring, tailoring — reads this parser’s output. Its errors do not stay local.
Read moreAI Engineering4 min
A dashboard that answers one question
A job seeker opens a dashboard to find out what to do today. Almost every dashboard answers a different question.
Read moreAI & Careers5 min
What automation gives and costs
Automation moves effort from the mechanical parts of a job search to the parts that matter. That is the honest claim; most others are overstated.
Read moreAI & Careers5 min
Suggesting or acting?
The difference is not capability. It is who is accountable for what goes out under the candidate’s name.
Read moreAI Engineering4 min
Filtering out what is not worth it
A large fraction of job search effort goes into applications that were never going to produce anything, for reasons visible in the posting.
Read moreAI Engineering5 min
Getting better, honestly
Improvement over time is a claim worth checking. Most systems that make it are just getting narrower.
Read moreAI Engineering5 min
Learning from a signal that barely exists
A rejection tells you almost nothing about why. Systems that claim otherwise are inventing explanations from noise.
Read moreAI Engineering5 min
A copilot, not an autopilot
The copilot pattern trades autonomy for trust, and in a job search that is usually the right trade.
Read moreAI Engineering5 min
Deciding what to do first
A job seeker has a few hours a week. Prioritisation is the feature that decides whether those hours were well spent.
Read moreAI & Careers5 min
Which produces better outcomes?
These are not competing strategies so much as different tools, and using one for everything is the actual mistake.
Read moreAI & Careers5 min
Technically yes, usefully no
The interesting question is not whether a machine can submit a hundred forms. It is what happens to the person whose name is on them.
Read moreAI & Careers5 min
Where personalisation runs out
It can produce a hundred different documents. Whether any of them are personalised is a different question.
Read moreAI & Careers5 min
The whole workflow, end to end
Automation covers the first half of a job search well and the second half badly, and the boundary is sharper than most tools admit.
Read moreAI & Careers5 min
Where the hours actually go
Most job search time is not spent writing applications. It is spent on things nobody would describe as valuable.
Read moreAI & Careers5 min
The right number, and why
Automation makes volume cheap, which makes the wrong strategy easy to pursue at speed.
Read moreAI & Careers5 min
A practical automation plan
Automate in the order of least risk and most tedium, and you will get most of the benefit before you reach anything that could embarrass you.
Read morePut the advice to work
Reading about ATS screening helps. Running your actual resume through one helps more.
New guides added regularly