SingularityX is a builder organization focused on AI-native financial systems, trusted execution, proof of contribution, and fair value distribution.
Our goal is to make every valid contribution recordable, verifiable, and assessable, and to distribute value fairly under clear rules.
We want to meet people who can form independent judgments and turn ideas into working results. Whether your strength is quantitative research, data engineering, backtesting and execution, AI and model evaluation, risk control, or contribution governance, you can find problems here that are worth sustained effort.
| Apply directly | Start with a public challenge |
|---|---|
| Send a résumé, GitHub profile, representative project, or research report. After mutual discussion and review, responsibilities, compensation, and contribution incentives are set out in formal agreements. | Choose a real challenge and submit reproducible work. Recognized work may earn the corresponding bounty and serve as evidence when applying to join. |
Want to build with us? Email your materials to
join@singularityx.techwith the subject:[Join SingularityX] <Name or GitHub Username>.
| ID | Challenge | Bounty cap | Current grades | Entry |
|---|---|---|---|---|
| SX-CH-001 | Multi-Hop Causal Chain Uncertainty Quantification | RMB 7,000 | S open; A awarded and closed | View challenge |
| SX-CH-002 | Tardis and Binance Market-Data Consistency | RMB 8,800 | S / A open | View challenge |
| SX-CH-003 | Runtime Contract from Paper Strategy to Real Backtesting | RMB 9,500 | S / A open | View challenge |
| SX-CH-004 | Robust High-Sharpe Strategy Mining from a Public Strategy Pool | RMB 12,000 | S / A open | View challenge |
Bounty amounts are public caps for complete challenges, not guaranteed payments. Recognition depends on the challenge requirements, reproducibility results, review conclusions, contribution evidence, applicable agreements, and compliance requirements. Use the status page as the authoritative entry point for grade changes.
We have received code submissions for multiple public challenges. After code review, reproducibility checks, and comprehensive assessment, only Challenge 1 (SX-CH-001) currently has submissions that meet the recognition standard; no submission for the other challenges has yet been recognized.
Challenge 1 has completed two review phases:
| Phase | Recognition | Current status |
|---|---|---|
| Phase 1 | One Grade B and one Grade C | Recognition and settlement completed |
| Phase 2 | One Grade A | Recognition and settlement completed; Grade A closed |
Public results retain only anonymous IDs, recognized grades, bounty status, and anonymized summaries. Authorized GitHub profiles and code completed a one-week public display and are no longer publicly accessible.
- Read the challenge board, the selected challenge, and the latest grade status.
- Prepare a complete code package, running instructions, reproduction commands, test results, environment details, known limitations, and rights statements.
- Submit privately by default to
join@singularityx.techwith the subject:[Challenge Submission] <Challenge ID> - <Name or GitHub Username>. - Use a public Pull Request only when intentionally disclosing the code and report. Never publish credentials, private data, account information, or non-public strategies.
Each contributor has one initial submission opportunity per challenge. An initial Grade A certification unlocks one complete Grade S-only upgrade within five calendar days (120 hours) after the confirmation email is sent, for a maximum of two submissions per contributor per challenge. See the contribution guide and the selected challenge for complete rules.
| Trusted research and engineering | Proof of contribution and governance |
|---|---|
| Data pipelines, backtesting, strategy research, model evaluation, risk controls, execution, and monitoring | Evidence chains, version history, reproducibility review, value attribution, appeals, and audits |
Our baseline principles are to open tasks while isolating control, reward contribution without accepting black boxes, and allow AI assistance without giving AI final authority. Actions involving funds, live execution, core accounts, or sensitive systems require separated permissions and human review.
We want long-term contribution inside the organization to be judged by verifiable evidence—not by title, presentation skills, or a single outcome. We continuously record what problem a person solved, how much quality improved, what risk they carried, and whether the work was actually adopted and reused. Research, validation, code, data, risk control, documentation, audits, and collaboration improvements can all constitute valid contribution.
| Dimension | What we examine |
|---|---|
| Validity | Whether the contribution addresses a real task, is adopted, and produces verifiable results. |
| Quality and trust | Whether the result is reproducible, the evidence is complete, the method is robust, and limitations are disclosed. |
| Marginal contribution | How much value the work adds—or how much loss or uncertainty it removes—compared with the counterfactual without it. |
| Long-term impact | Whether the result can be reused, extended, and retained as an organizational asset that continues to create value. |
| Risk and collaboration | Whether the work respects permission, compliance, and risk discipline and improves the team’s overall delivery quality. |
Assessment follows a contribution-record, evidence-verification, value-attribution, and human-review process. Self-reported claims may provide leads but do not receive full weight by themselves; material conclusions require support from code, data, experiments, review records, or evidence of actual adoption. We assess both final outcomes and the research, validation, and risk-control work that made them possible.
- Define the distribution boundary first: formal agreements specify the value pool, eligible participants, settlement conditions, and exclusions. Public-challenge bounties and long-term member contribution are administered separately.
- Use multiple attribution anchors: direct contribution, marginal contribution, counterfactual necessity, reuse impact, and collaboration stability are considered together; hours, title, or a single score do not determine the result by themselves.
- Preserve the total: contribution shares are normalized and checked against ranges and caps. Distribution cannot exceed the confirmed value pool, and the same outcome cannot be counted twice.
- Keep review and appeal available: material human adjustments require recorded reasons and conflict-of-interest disclosure. Evidence, rules, and outcomes remain traceable, and members may request review within the applicable window.
The purpose is not to replace judgment with one score, but to make judgment evidence-based, distribution bounded, adjustments recorded, and outcomes explainable. See the Value Assessment System and Value Distribution Protocol for the detailed framework.
All compensation, bounties, profit sharing, equity, options, tokens, points, or other economic interests are governed by formal agreements, applicable law, tax requirements, and organizational approval.
| Document | Scope |
|---|---|
| Full User Agreement, Challenge Rules, and Legal Statement | Participation, review, settlement, intellectual property, and disputes |
| Privacy Policy | Information collection, use, disclosure, retention, and rights |
| Important Legal and Compliance Statement | Research boundaries, risk, and compliance notices |
| Difficulty and Bounty Amounts | Difficulty levels, bounty caps, and recognition scope |
| Challenge Source Integrity | Challenge-file hashes and verification |
Platform content, tasks, examples, models, code, backtests, and user work are for lawful technical research and engineering validation only. They do not constitute investment advice, trading signals, return promises, asset management, or financial-product sales.
If you believe in trusted execution, proof of contribution, and fair distribution—and want to turn research into systems that work—apply to join SingularityX.
Please include a résumé, GitHub profile, representative project, or research report.
- Membership applications:
join@singularityx.tech, subject[Join SingularityX] <Name or GitHub Username> - Challenge submissions, privacy requests, and compliance complaints:
join@singularityx.tech - Rule version: V2.1.0 (Second Version; updated 2026-07-21)
- Rule interpretation body: SingularityX