Pillar 01
CI & Agent Reliability
Diagnose CI failures, bound autonomous repair, and govern AI agent actions with policy gates, human escalation, and audit logs.
Taiwan-based · Accounting + systems background
Zixsa
CI & Agent Reliability for Financial Systems · Quant Trading Infrastructure
I build systems that turn CI failures, AI agent actions, and financial data into auditable evidence, bounded automation, and decision-ready controls.
About
An auditor's instincts, applied to engineering
I come from accounting and compliance-oriented work: tax administration, SAP system support, and regulated workflows where every record has to hold up to review. I now build software with the same instincts: explicit controls, separation of duties, and evidence you can audit.
My public work applies that lens to three areas: keeping CI pipelines and AI coding agents inside bounded, reviewable automation; building financial data and decision systems that fail closed; and quant research infrastructure that rejects leakage and overfitting before capital is at risk.
Technology Risk Lens
01
Fail closed
Automation defaults to reject or escalate. Nothing is approved silently.
02
Evidence over claims
Every run leaves an append-only, replayable audit trail.
03
Separation of duties
The component that evaluates a change is never the one that approves it.
04
Honest scope
Each repository states what is implemented, what is simulated, and what is out of scope.
What I Build
Three pillars, one control mindset
Pillar 02
Financial Systems & Tech Risk
FinOps controls, financial data quality, and valuation workflows designed the way an auditor would review them.
Core Specialty
Quant Trading Infrastructure
Point-in-time data, purged walk-forward validation, cost-aware backtests, and execution controls, backed by freelance multi-exchange work since 2021.
Featured Work
Systems with evidence attached
Experience
Execution across regulated, technical, and live operating environments
Freelance Quantitative Trading Systems Developer
2021-Present
- Designed Python and Go trading systems with automated execution pipelines.
- Built event-driven architectures for real-time market data processing.
- Integrated exchange APIs including Binance, KuCoin, and Gate.io.
- Implemented risk control frameworks and portfolio allocation models.
- Conducted backtesting using time-series data with Sharpe ratio, drawdown, and win-rate analysis.
Information Systems Project Trainee
iSpan, 2023
- Developed web-based system modules including authentication and notification flows.
- Designed structured workflows and testing processes for predictable delivery.
- Improved reliability and data integrity through consistent validation and QA practices.
IT Support
Jetwell, 2023
- Assisted SAP system deployment and virtual machine setup.
- Supported test execution and operational environment preparation.
- Helped stabilize rollout activities with practical system support.
Intern
National Taxation Bureau, 2019
- Verified tax data, documentation, and administrative records.
- Worked within strict compliance, auditability, and internal control systems.
- Developed discipline for precision in regulated workflows.
Live Streamer
17LIVE, 2025-Present
- Managed content operations under platform compliance rules.
- Handled real-time interaction and decision-making in live environments.
- Built consistency in performance-driven settings with immediate feedback loops.
Skills
Capabilities across code, controls, and markets
CI & Agent Reliability
GitHub Actions, CI/CD failure triage, policy gates, retry budgets, audit logging
Automation that stays bounded, reviewable, and reversible.
Financial Systems & Tech Risk
Internal controls, segregation of duties, FinOps (FOCUS), data quality, DCF analysis
Financial workflows designed to survive audit, not just produce numbers.
Quant & Data
Walk-forward validation, cost-aware backtesting, time-series analysis, risk metrics
Sharpe, VaR, Expected Shortfall, and drawdown treated as evidence, not marketing.
Engineering Stack
Python, Go, TypeScript, SQL, FastAPI, Docker, Linux
Event-driven services, APIs, and tooling that run the same locally and in CI.
Next Layer
Review the evidence behind each system.
Every project lists the problem, how the system works, and what you can verify in the repository.