FiClaw / Private securities funds and asset managers
AI research workflows for private funds
For teams researching A-share strategies, aligning experimental assumptions or improving research handoffs. First establish whether one real question can be specified, implemented and reviewed by another researcher, then assess organizational collaboration and integration needs.
Published by FiClaw 产品团队 · Published · Updated
阅读中文版Where FiClaw fits
FiClaw is an AI-native quant strategy research workstation developed by SiStarAI. Private securities funds and asset managers can start with a scoped research task, turn an idea into a specification, and connect code generation, A-share historical backtesting, diagnosis and parameter comparison. Researchers and reviewers check the data, implementation and experimental evidence before deciding what follows.
Start with a concrete research task
01
Specify an investment hypothesis
- Input
- A researcher supplies the universe, signals, holdings and rebalance conditions.
- Research output / evaluation scope
- Use FiClaw to help organize rules, open questions and a Python implementation.
- Human review
- The research lead confirms information timing and trading behavior; runnable code alone is insufficient.
02
Compare experiments consistently
- Input
- Fix the historical window, benchmark, costs and candidate parameter ranges.
- Research output / evaluation scope
- Connect historical runs, diagnosis and parameter comparison, then document findings.
- Human review
- Review data coverage, costs and out-of-sample evidence; retain failed and inconclusive experiments.
03
Prepare a research handoff
- Input
- Collect the specification, code, settings and outputs for the same question.
- Research output / evaluation scope
- Organize review materials against an agreed checklist, including changes and unresolved issues.
- Human review
- Another researcher checks rerun feasibility. Central archiving and organizational permissions require pilot scoping.
A pilot with acceptance criteria
- 01
Choose one experiment
Select an A-share question with accessible data and explicit rules. Agree on permitted inputs, outputs and accountable people.
- 02
Establish a manual baseline
Use the same data scope and assumptions, recording preparation, coding, debugging and review time for comparison.
- 03
Review the evidence
Check specification fidelity, explainability and visible exceptions. Include human corrections and model usage in total cost.
- 04
Decide whether to expand
Only after agreed acceptance criteria are met, discuss additional users, internal interfaces, governance and deployment.
Confirm capabilities and delivery scope
- Current public capabilities
- The public entry point is Desktop with cloud quant research services: strategy design, code generation, A-share historical backtesting, diagnosis and parameter comparison. Check the coverage page and actual API responses for data availability.
- Separate assessment required
- Shared institutional agents, centralized permissions and audit, internal interfaces, and VPC, hybrid or private deployment require separate assessment of data flows, interfaces and operations. The public installer does not establish that these capabilities have been delivered.
- Decisions and accountability
- FiClaw supports research. It does not take investment decisions, approve compliance or place live orders; authorized institutional staff remain responsible for conclusions, limits and deployment decisions.
Procurement and pilot questions
Is this for private securities funds or private equity?
This workflow primarily addresses quantitative strategy research at private securities funds and asset managers. Public backtesting focuses on A-shares. Private equity valuation, private asset due diligence and other asset classes require separate data and workflow assessment.
Can a private fund team try FiClaw directly?
Download Desktop from the official website and evaluate one scoped research task using permitted data. Confirm the pilot scope and data handling before adding multiple users, sensitive data or internal systems.
Does Desktop guarantee private deployment or data residency?
No. A desktop interface does not mean model calls, data requests and backtest computation all stay local. Data flows, storage, service dependencies and operational responsibilities need explicit assessment.
How should a team decide whether to expand a pilot?
Compare completion time, review outcomes, rerun consistency, human corrections and total cost on the same task. Historical returns alone are insufficient acceptance criteria for a research tool.
Further reading
Private fund AI pilot acceptance checklist ↗
Define the baseline, deliverables, reviewers and stop conditions.
Backtest report review ↗
Check the evidence connecting data, metrics and conclusions.
A-share data coverage and limitations ↗
Verify the universe, historical window and actual data availability.
Bring one research question
Describe the team roles, task, available material and expected output to start a scope discussion. Public or redacted samples are enough for an initial conversation.