loading ...
new
now with built-in ats.
sign up freehow it works
onecandidate.
every signal, read.
post a role and step back. aperture runs the interview, reads how each candidate thinks, confirms who actually showed up, and hands your team a ranked shortlist with the reasons behind it.
one candidate, end to end
post the role. read the shortlist.
everything between those two moments runs on its own. here is what happens to a single candidate, from the day they apply to the day you decide.
01
they apply
the role goes up. every applicant lands in one place.
02
aperture interviews
a real, adaptive conversation, on the candidate's schedule.
03
it reads how they think
voice, face, and language, captured as one signal.
04
it verifies who showed up
location, device, and network, confirmed the moment it begins.
05
it scores the evidence
six dimensions, each with honest uncertainty.
06
it ranks the pool
every candidate placed against the whole cohort.
07
you decide
a shortlist with the reasons already attached.
the pipeline
where aperture fits your stack.
job board / ats
linkedin · greenhouse · ashby · lever
aperture intelligence
aperture
behavioral assessment · scoring · ranking
your hiring team
final interviews · decisions · offers
job board / ats
linkedin · greenhouse · ashby · lever
aperture intelligence
aperture
behavioral assessment · scoring · ranking
your hiring team
final interviews · decisions · offers
the interview
a structured interview, that adapts to each candidate.
click through each stage to see what happens between an application arriving and a score being assigned.
step 01
role context ingested
aperture reads the job description, required skills, and any role-specific signals you define. the interview adapts its question bank to the role.
NeuralPrint axon
it reads how they think, not just what they say.
most tools score the words. NeuralPrint reads the reasoning underneath them, the way someone frames a problem, where they hesitate, how they recover, and turns it into a structured read no resume can fake.
12k
signals per interview
collapse into one profile.
a structured map of how a person thinks, the same evidence base for every candidate, ready the moment the interview ends.
go inside NeuralPrintevery interview, verified the moment it begins.
where they really are. the moment an interview begins, aperture resolves the candidate's true location from their ip and cross-checks it against their device clock. a geo-fence mismatch surfaces instantly.
the device they're really on. operating system, browser, display, and camera are fingerprinted into a single signature. the same person interviewing twice from two machines never goes unnoticed.
how they're really connecting. vpns, proxies, datacenter origins, and automation flags are checked live against blocklists. an interview routed through an anonymizer is flagged, not trusted.
sealed, end to end. every signal is captured client-side and bound to the session over an encrypted channel. one complete, tamper-evident record of who actually showed up.
the verification layer
every interview, verified the moment it begins.
1 in 3 hiring managers caught a fake identity in an interview last year. aperture confirms where a candidate is, what they're on, and how they connect, before a single answer is scored.
live session7f3a · capturing
locationresolving
Cupertino, California, United States
37.323° N · 122.032° W
residential networkgeo-fence·tz mismatch
network intel
your ip
76.102.183.136
ipv4 · resolved live
asn
AS6167
routing · verified
isp
Verizon Fios
residential broadband
classification
residential network
no anonymizer detected
geo source
ip + browser geo
cross-checked · mismatch
ip timezone
Los Angeles
device says UTC
encryption
TLS 1.3 · sealed
aead · verified handshake
session integrity
session 7f3a
sha-256 · device-bound
device fingerprint
operating system
resolving
-
browser
resolving
-
display
resolving
-
camera
resolving
-
session environment
vpnclear
proxyclear
datacenter originclear
residential ispverified
location verifiedtz mismatch
automation flagclear
extended displayclear
ip versionIPv4
aperture · session sealedjun 25, 12:26 pm
scored · λ-core · σ-10
six dimensions, with honest uncertainty.
every interview is scored across six behavioral dimensions on a 10-point scale. each score comes with an 80% confidence interval, the range captures genuine uncertainty rather than false precision.
statistically indistinguishable candidates are flagged. rank 9 vs. rank 11 is often noise, not signal.
COGcognitive reasoning
DOMdomain knowledge
COMcommunication
BEHbehavioral indicators
COLcollaboration
ADPadaptability
λ-core · σ-10 · candidate #247
dimbars scoreconfidence intervalvalreliability
COG6.4
COG
5.96.8
6.4moderate
DOM5.8
DOM
5.26.3
5.8low
COM6.7
COM
6.37
6.7moderate
BEH6.1
BEH
5.66.5
6.1low
COL5.3
COL
4.85.7
5.3low
ADP6.5
ADP
66.9
6.5low
σ-10 composite
6.21 [5.82, 6.54]P(top 5%) = 91.2%
ranked · your decision
ranked against the cohort. the shortlist earns its place.
every candidate is positioned relative to everyone else who interviewed for the same role. as more people interview, the rankings sharpen. what your team reads is an ordered argument, not a list.
the score updates the moment another candidate is interviewed, so you always know who is actually ahead, with the evidence behind every placement.
the science
why it actually works.
behavior
hover a principle to watch it resolve.
get started today
human intelligence,
on your side.
switch it on today. tomorrow you read a shortlist, not a pile.
free to start, no credit card, set up in minutes