Great candidates reveal themselves earlier than most companies realise.
Kristina Golovko
MindDesign
The signal-first system
Signal
Observable output before application
Behaviour
How they engage with hard problems
Capability
What the signals predict
Hiring decision
Grounded in evidence, not presentation
CV-first hiring surfaces the best CV writers. Signal-first hiring surfaces the best candidates.
Why this matters
A CV is a marketing document. It is designed to present the candidate's experience in the most favourable light possible. It tells you what someone has done — not how they think, how they solve problems, or how they'll perform in the role you're hiring for.
The signals that actually predict capability exist elsewhere: in code, in writing, in how they ask questions, in the problems they choose to work on, in the communities they contribute to. These signals appear long before a CV is ever formatted.
Signal-first sourcing shifts evaluation earlier in the process — surfacing predictive evidence before the formal hiring timeline begins. The result is a more accurate pipeline, better interview questions, and more confident decisions.
Founder reality
Build a signal inventory before the next search opens:
What would a highly capable person in this role have built, written, or contributed to that we could find before they apply?
Which communities or platforms surface the output of strong candidates in this domain?
What does the quality of someone's questions reveal about the quality of their thinking?
Are we using CV-based filters that systematically exclude strong non-traditional candidates?
When we review CVs, are we looking for capability indicators — or just for familiar patterns?
Signal inventories take 30 minutes to build. They permanently change the quality of candidates who enter the process.
The framework
Each signal type is observable before formal application. Build sourcing logic that reaches them.
Output signals — code, writing, systems, and artifacts they've created
GitHub repositories, open-source contributions, technical blog posts, papers, design documents, side projects. Output signals reveal how someone thinks, what problems they find interesting, and how they execute. Review output before reviewing the CV for technical roles.
Engagement signals — how they participate in professional communities
The quality of how someone engages in forums, Slack communities, open-source issues, or conference Q&As reveals thinking patterns that interviews struggle to surface. A candidate who consistently reframes problems before answering them signals something specific about how they reason.
Question quality — what they ask reveals what they understand
Strong candidates ask better questions than weak ones. In technical interviews, on community forums, in outreach responses. The questions they ask — their specificity, their depth, their direction — signal the map they're working from.
Trajectory signals — how they've developed and where they're heading
Pattern of development over time is a strong predictor of future trajectory. A candidate whose work shows meaningful progression in depth and complexity over 3–5 years is a different hire from one whose profile has been static. Trajectory signals require looking across time, not just at the current state.
Common mistakes
01
Requiring a CV before reviewing signal
For technical roles, requiring a CV as the first step filters out excellent candidates who haven't maintained traditional presentation. Reverse the order: signal first, CV for context.
02
Treating all signals as equal
Output signals carry more predictive weight than engagement signals for most technical roles. Calibrate signal weight to the specific capabilities the role requires.
03
Not using pre-interview signals to change the interview
Signal review that doesn't change the interview questions is wasted. If a candidate has written specifically about the technical domain, the interview should engage with that directly.
04
Ignoring signals from non-traditional backgrounds
Signal-first evaluation should actively include candidates whose CVs would be screened out by traditional filters: non-traditional education, non-linear career paths, self-taught backgrounds. The signal matters — not the path that produced it.
Example scenario
A language model infrastructure startup. Hiring a senior ML systems engineer. CV-first screening for 6 weeks: 3 qualified candidates, none progressed to offer.
The signal-first switch
Sourcing shifted to output discovery: GitHub contributors on key ML systems repositories.
Technical blog: authors of posts on inference optimisation, quantisation, and serving infrastructure in the past 12 months.
Community: participants in ML systems Discord and Slack communities who asked and answered technically substantive questions.
CVs reviewed after signal identified — not as the first filter.
What signal-first revealed
One candidate had no university degree listed on LinkedIn. GitHub showed 3 years of consistent, architecturally significant contributions to a major inference serving project. Technical blog contained 4 posts demonstrating deep understanding of the exact system they'd be hired to build.
The outcome
Candidate joined the process. Offer extended and accepted in week 4 of the signal-first search. CV-first filter would have screened them out before the first call.
CVs filter for presentation. Signals filter for capability.
The best hiring decisions are made by people who learned to read what candidates reveal before the formal process begins. Build the signal inventory. Use it first.
Related thinking
Precision Sourcing
What Great Engineers Signal Early
Strong engineers leave clues before interviews.
Precision Sourcing
Precision Sourcing Explained
Better hiring starts with better search logic.
Precision Sourcing
Why Strong Candidates Stay Invisible
Great talent is often hidden in plain sight.
Continue reading
Back to all Playbooks →