Trading agents · Round 1 result
Arnav Chauhan won with +5.91%.
QQQ ended the same window at -3.9%. Arnav's code is public and now runs as the benchmark every Round 2 entrant must beat.
Crowdsource AI solutions. Pay for verified outcomes.
Bring one measurable problem. We build the test, builders compete under published rules and scoring, and you pay only when a solution clears the agreed bar.
3 challenges open · $3,800 in prizes · public results · verified winner audits
Public challenge boards
Builders are proving real solutions in public.
How it works
From problem to verified result.
01
Define the result
Tell us the input, the useful output, and what solving it is worth.
02
Run one shared test
We agree the data, scoring, qualifying bar, and practical limits before launch.
03
Publish what worked
Builders compete on the same evaluation. Only work that clears the bar can win.
If nothing clears the bar, there is no forced winner. The bounty rolls over or comes back.
For problem owners
A real business problem in. A verified answer out.
Bring a workflow, footage, or dataset that matters. Builders return an answer, evidence, and run cost against the same test. Kitchen CCTV is one live example.
Challenges
Live challenges, past results, and what's next.
Enter a live contest, inspect a finished result, or help shape the data for the next one.
Kitchen CCTV monitor
kitchen CCTVFor restaurants and cloud kitchens: answer hidden questions from CCTV-style footage about caps, handoff, bottlenecks, visible steps, and what was not visible.
4 scored entries
Prize$300 to the winnerTo qualify$0.30 / 60 min capTested onHidden questionsClosesSep 9
View challenge →Trading agents
trading agentsWrite one Python function that beats Arnav on a fresh live market window — no finance background, no money, no API key.
Live participant board
Prize$1,000 to top 3To qualifyLive market scoreTested onLive market dataClosesSep 4
View challenge →Signalpost company research
company researchTurn official filings and public sources into verified profiles for every Norway-registered company—and keep them current.
$2,500 prize · partner with Håvard to launch in Norway
Prize$2,500 final rewardsTo qualify65/100 + recall/precision gatesTested on100 random companies per dayClosesOct 21
View challenge →Local dictation · shipped as RambleFix
Free, local Hindi+English dictation—built for reliable bilingual speech.
RambleFix types English and mixed Hindi+English at the cursor, runs locally on your Mac, and costs nothing. It was built specifically for code-switched speech.
Early briefs for the next challenges.
We are shaping the test data and scoring now. Share representative data or register interest; the bounty and rules will be published before entries open.
Sports automated scoring
Early briefTurn sports footage or event feeds into accurate scores and game-state updates.
What we need: Sample footage or an event feed, plus trusted final scores and game states.
Bounty and scoring before launchRegister interest →
Opinion markets
Early briefEstimate the odds of real-world events, explain the signal, and trade against a fixed benchmark.
What we need: Real questions, source odds, decision dates, and unambiguous outcomes.
Bounty and scoring before launchRegister interest →
Operations reconciliation
Early briefMatch purchase orders, invoices, and delivery records, then surface the exceptions that need action.
What we need: Anonymized purchase orders, invoices, delivery records, and resolved exceptions.
Bounty and scoring before launchRegister interest →
Have a problem? Post it as a challenge and we will turn it into a scored test.
Live boards
Reward
$300top valid score
Cost cap
$0.30 / 60 min capnormalized per 60 minutes
Round 1 ends
Sep 915 days left
For restaurants and cloud kitchens: answer hidden questions from CCTV-style footage about caps, handoff, bottlenecks, visible steps, and what was not visible.
6 submissions reviewed; 4 have verified scores. Answer accuracy ranks first, then lower model/API cost.
Kitchen CCTV monitor · live$0.30 / 60 min cap
Public samples
60 min + CCTV clips
Scored on
Hidden questions
Entries
6 reviewed
| # | Entry | Answers correct | Model cost / 60 min |
|---|---|---|---|
| 1 | Kim Beomjin · verified runAug 26 rerun on commit df0f143 completed cleanly in 43 seconds. It used 11 frames and 6 local model calls; the person count at 00:45 and island interaction at 15:00 remain the misses. | 4 / 6 | $0.00 |
| 2 | Vishal · verified runThe Aug 26 rerun on commit 546bca3 improved to 4 / 6. It completed cleanly in 4m 59s at $0.00; head-cover counting and the 00:45 island-contact check remain the misses. | 4 / 6 | $0.00 |
| 3 | Meet · verified runTied on answers; the Aug 10 cloud run stayed under the cap at $0.0312 per 60 minutes. | 3 / 6 | $0.0312 |
| 4 | Raam · verified runAug 20 rerun on commit 2854a8c completed cleanly against the official six-question set at $0.00. The pipeline returned valid answers and evidence; person counts and action detection remain the main gaps. | 2 / 6 | $0.00 |
1Kim Beomjin · verified run
4 / 6 correct$0.00 / 60 min
Aug 26 rerun on commit df0f143 completed cleanly in 43 seconds. It used 11 frames and 6 local model calls; the person count at 00:45 and island interaction at 15:00 remain the misses.
2Vishal · verified run
4 / 6 correct$0.00 / 60 min
The Aug 26 rerun on commit 546bca3 improved to 4 / 6. It completed cleanly in 4m 59s at $0.00; head-cover counting and the 00:45 island-contact check remain the misses.
3Meet · verified run
3 / 6 correct$0.0312 / 60 min
Tied on answers; the Aug 10 cloud run stayed under the cap at $0.0312 per 60 minutes.
4Raam · verified run
2 / 6 correct$0.00 / 60 min
Aug 20 rerun on commit 2854a8c completed cleanly against the official six-question set at $0.00. The pipeline returned valid answers and evidence; person counts and action detection remain the main gaps.
Reviewed, not ranked
No score is published until setup and integrity checks pass.
AryaRevise and resubmit
The Aug 26 rerun on commit b501a60 completed within the frame and runtime caps, but benchmark-specific timestamp answer overrides failed the generalization check. No verified score is published.
SankeerthSetup incomplete
Commit eebd656 compiled, but the evaluator could not complete the declared 4B model setup, so no official answer file or score was produced. A reproducible model warm-up path is required.
Watch · 2 minutes
If you can score it, you can post it.
A problem you keep meaning to deal with — and the move most people don't have a habit for yet: turn it into a challenge, let builders compete on the same test, and give qualifying builders a public proof of what they built.
After a published result, builders can claim their profile. Qualified builders who want a signed certificate can share that profile on LinkedIn with what they built and liked or learned; sharing never changes the score. Browse public results →
Public proof
Winning bots clear the goal, not just the demo.
These are public examples of builders meeting a measurable bar. We publish the shared inputs, the score, the limits, and the work so the result can be checked.
Trading · published graph
Market down. Builders held. Market up. Builders competed.
Diagnostic graph from the Jul 7–27 live window; the round remained open.
Open the winner project →Local dictation · shipped outcome
Free, local Hindi+English dictation—built for reliable bilingual speech.
Sankeerth's winning approach shaped how RambleFix remembers important words. Arnav's work helped improve product and technical terms in Hindi+English mode.
RambleFix types English and mixed Hindi+English at the cursor, keeps audio on your Mac, and costs nothing. It was built specifically for code-switched speech—not just a generic multilingual language list.
See the builders, code, and measured results behind the public challenges.
Browse public results →Pay for a verified outcome, not another demo.
Bring one AI or automation problem with an answer we can measure. We define the test, recruit builders, and run the competition. Here are a few examples:
- Local dictation — a local, private Wispr Flow-style tool that handles mixed-language speech.
- Hotel / Airbnb housekeeping QC — was the room cleaned to standard? Photo in, pass/fail out.
- Kitchen CCTV monitor — use existing restaurant CCTV to check caps, hygiene steps, handoff delays, and bottlenecks.
- Cart-abandonment nudges — an agent watches a checkout and decides if and when to step in.
If you can describe what goes in and what good looks like, we can usually turn it into a test in about a week, then run a few weeks of competition. The bounty is paid only when a solution clears the bar you approved.
How much should the bounty be? Size it to the problem. A simple way to think about it: if cracking it is worth a builder's weekend, the prize should feel worth that weekend. Around $1,000+ can work for a focused task; larger pools pull in stronger fields.
Send your problem →Who's building this
Two friends — product-and-tech geeks, both ex-founders — who hit this exact problem in our own work every week: you need an agent for something that matters, and no honest way to tell which approach is actually best. builderr is our fix.
We put real money on it. Each challenge has its own sponsor and a published test. We show what was run, what cleared the bar, and what builders can inspect afterward. No black box: the point is proof, not promises.
Current challenges have no platform fee. The bounty passes through to qualifying winners, minus agreed running costs. If no solution clears the bar, there is no forced winner. Have a problem we can score? Get in touch.
builderr.ai — built by two ex-founders who got tired of guessing.how to enterfor agentsrules & faqgithubinquiries@builderr.ai