START Warsaw

Partner talent brief

Builders who ship, shown through the work.

An evidence-bound view of product seriousness, technical substance, originality, execution, and validation across the applicant community.

38 of 286 (13%)advanced or exceptional technical depth
80 of 286 (28%)differentiated problem framing
17 of 286 (6%)primary or end-to-end execution
17 of 286 (6%)explicit external validation observed
Event OpenAI HackathonEvent date 2026-07-11Scope Partner-safe evidence summary

Interactive aggregate view

Choose a cohort, then inspect the evidence

The cohort control changes every count, share, denominator, and comparison. The metric groups stay fixed so the evidence remains comparable. Select or focus a metric for its definition and limits. No score or ranking is created.

Cohort

All applicants

286people in the selected cohort

All valid applicants in the reviewed event population.

Project evidence

Comparable signals about work already built, technical substance, and problem framing.

Demonstrated capabilities

Overlapping technical capability signals from reviewed work; these counts are not additive.

Career delivery

Overlapping evidence of shipping, founding, and customer delivery; this is not a seniority ladder.

Selected evidence

Prototype or beyond

Prototype, working-product, or production maturity.

Evidence standard
Reviewed application, public-project, and event-submission evidence; missing evidence remains unknown.
Selected denominator
286
Evidence limits
Count withheld by the small-group privacy rule. Not every reviewed source can support every metric. No recorded positive evidence for this signal means unknown, not a negative assessment.

Compare the same signal

All applicants187 of 28665% of cohort
Accepted participants67 of 8381% of cohort
Confirmed attendees59 of 7876% of cohort
Cross-dimensional evidenceFull applicant pool · denominator 286

Each applicant appears in exactly one intersection row. The eight rows partition all 286 applicants. Filled dots mean the application signal was recorded; empty dots mean signal not recorded. Signal not recorded is not a negative assessment.

Founder / co-founder evidence53 of 286
Technical function213 of 286
Shipped-product evidence84 of 286
FounderTechnicalShippedExact intersectionPeopleShare

Insufficient founder context: 209 · Unclassified role or stage: 133 · Unclassified function: 55 · Insufficient shipping context: 60

Technical only · 125

Exact application-evidence cell. Recorded: Technical function. Signal not recorded: Founder / co-founder evidence, Shipped-product evidence. A signal not recorded is not a negative assessment.

Reviewed application evidence, not a quality score.
Download PDF

How much demand converted into participation?

01

Demand and participation

One people-based story. Width shows volume from application to reviewed attendance.

01Applied286 people

Full applicant pool

02Accepted by organizers83 people

29% of applicants

03Confirmed present at the event78 people

78 of 83 accepted participants were confirmed present; 5 were not in the reviewed attendance count.

Applicant starting pointApplication-reported team status
  • Applied without a pre-formed team128 of 286 (45%)
  • Applied with a pre-formed team158 of 286 (55%)

What have applicants actually built?

02

What people have actually built

A protected cohort matrix compares positive evidence and denominators without treating selection as quality.

Comparable cohort signalsDescriptive shares, not a selection score
All applicantsDenominator 286 · Unknown withheld
Accepted participantsDenominator 83 · Unknown withheld
Confirmed attendeesDenominator 78 · Unknown withheld
Prototype or beyond1
187 of 286All applicants · 65% of cohort
67 of 83Accepted participants · 81% of cohort
59 of 78Confirmed attendees · 76% of cohort
Advanced or exceptional technical depth2
38 of 286All applicants · 13% of cohort
22 of 83Accepted participants · 27% of cohort
17 of 78Confirmed attendees · 22% of cohort
Differentiated problem framing3
80 of 286All applicants · 28% of cohort
38 of 83Accepted participants · 46% of cohort
31 of 78Confirmed attendees · 40% of cohort
Moderate-or-stronger technical depth6
150 of 286All applicants · 52% of cohort
59 of 83Accepted participants · 71% of cohort
51 of 78Confirmed attendees · 65% of cohort
Positive evidence available for the full applicant poolNested cohort counts are withheld by the privacy rule. Missing evidence is not a negative assessment.
Primary or end-to-end execution417 of 286
Explicit external validation observed517 of 286

Public-project evidence used 172/286 applicants · 56/83 accepted participants · 50/78 confirmed attendees. Non-use is not a quality judgment.

Who is represented in the applicant community?

03

Who is in the applicant community

Student, seniority, founder, and technical-function context, with source boundaries and unknowns visible.

Application-reported primary role or stage

One mutually exclusive role or career-stage category per applicant; the distribution totals 100%, including unknown.

  • Founder / co-founder role53 of 286 (19%)
  • Junior20 of 286 (7%)
  • Lead, staff, principal, head, or executive17 of 286 (6%)
  • Mid-level10 of 286 (3%)
  • Senior13 of 286 (5%)
  • Student40 of 286 (14%)
  • Unknown / not classified133 of 286 (47%)

Reviewed founding evidence

Current and former founder states are mutually exclusive here; founder state may overlap the primary role or stage shown at left.

  1. Current founder22 of 286 (8%)
  2. Former founder7 of 286 (2%)

Career functions

Top functions shown. Functions supported by available career evidence; people may span several.

  1. Software engineering164 of 286 (57%)
  2. Data and AI117 of 286 (41%)
  3. Product29 of 286 (10%)
  4. Research29 of 286 (10%)
  5. Operations22 of 286 (8%)

Where do role, technical function, and shipped-product evidence intersect?

04

How applicant signals intersect

Exact reviewed application-evidence intersections across founder role, technical function, and shipped-product evidence.

Each applicant appears in exactly one intersection row. Filled dots mark recorded signals; empty dots mean signal not recorded, not a negative assessment.

Founder / co-founder evidence53 of 286
Technical function213 of 286
Shipped-product evidence84 of 286
FounderTechnicalShippedExact intersectionPeopleShare
Technical only12543.7%
Technical + shipped5318.5%
None recorded4816.8%
Founder + technical + shipped186.3%
Founder + technical175.9%
Founder only124.2%
Shipped only72.4%
Founder + shipped62.1%

Unclassified or insufficient context: Insufficient founder context: 209 · Unclassified role or stage: 133 · Unclassified function: 55 · Insufficient shipping context: 60.

Which capabilities and market contexts are represented?

05

What the community can build

Concrete methods and capabilities, followed by the market contexts visible in the reviewed work.

Technical methods

Technical approaches visible in the evidence-bound work; people may use several.

  1. Applied AI / ML111 of 286 (39%)
  2. Full-stack web82 of 286 (29%)
  3. Automation orchestration67 of 286 (23%)
  4. Data engineering53 of 286 (19%)
  5. Natural language processing51 of 286 (18%)
  6. Realtime systems21 of 286 (7%)

Demonstrated capabilities

Capabilities demonstrated by concrete work, not self-declared skill lists.

  1. Product engineering134 of 286 (47%)
  2. Data and AI engineering123 of 286 (43%)
  3. Backend engineering115 of 286 (40%)
  4. Frontend engineering78 of 286 (27%)
  5. Research experimentation36 of 286 (13%)
  6. Infrastructure / DevOps32 of 286 (11%)

Where the work is aimed

Top domains shown; reviewed market context; categories may overlap.

Market domains

Top domains shown. Market contexts explicitly supported by the available evidence; people may span several.

  1. Commerce consumer58 of 286 (20%)
  2. Developer infrastructure39 of 286 (14%)
  3. Enterprise operations34 of 286 (12%)
  4. Financial services27 of 286 (9%)
  5. Media creative24 of 286 (8%)
  6. Healthcare and life sciences17 of 286 (6%)

What evidence was available for this release?

06

What evidence was available

Reviewed source use, explicit non-use, and the boundary between coverage and quality.

Evidence source use

Assessment boundary. These counts mean a source was cited in a person-level evidence review, not attendance, submission completion, or a quality score. The partner output excludes names, contacts, profile links, and raw provider records.

  • Application270 of 286Application evidence cited in reviewed facts.
  • Public projects172 of 286Public-project evidence cited in reviewed facts; non-use is not a quality judgment.
  • Event submission61 of 286Submission evidence cited for reviewed people; this is not attendance or team completion.
  • Dedicated career-provider evidenceNot usedDedicated career-provider evidence was not used in this release.

How should partners interpret the evidence?

07

How to read the evidence

Definitions, denominators, unknown states, and privacy rules for every claim.

Evidence definitions

Prototype or beyond
Prototype, working-product, or production maturity.
Advanced or exceptional technical depth
Non-trivial architecture, systems constraints, or production methods are visible in the work; language and stack lists alone do not qualify.
Differentiated problem framing
The work addresses a specific non-routine problem with a distinct approach or constraint; topic keywords alone do not qualify.
Primary or end-to-end execution
Application or submitted-project evidence attributes primary or end-to-end delivery to the participant; repository ownership alone does not qualify.
Explicit external validation observed
Reviewed evidence explicitly names users, adoption, traction, or credible third-party recognition. Non-observation means the current sources did not support this positive claim; it is not a negative rating.
Moderate-or-stronger technical depth
Concrete implementation or systems evidence supports at least moderate technical depth; a stack list alone is insufficient.
Category overlap
Primary role or stage is exclusive. Founder state, career functions, capabilities, methods, and domains are independent overlays; a person may appear in several. Do not sum the overlays or add them to the 100% distribution.

Interpretation

Denominator
Evidence shares use the full eligible applicant population. Source-use coverage is reported separately and is not a quality score.
Unknown
Evidence was missing, insufficient, unavailable, or unresolved. Unknown never means a negative assessment.
Unclassified
The available inputs did not support one controlled career, role, capability, or domain category.
Not in this release
Gender, age, nationality, location, and graduation year were unavailable or unreviewed and were not inferred.
Whole-person unresolved
18 of 286 (6%) had missing or conflicting reviewed evidence. They remain outside positive claims; this is not a negative judgment.
Privacy threshold
Counts, complements, and related intersections under 5 are withheld.