CHRISTIANCHIBA
SYSTEMS BUILDER / AI-NATIVE ENGINEERING
Project-led systems builder focused on quantitative platforms, agent orchestration and owner-operated infrastructure. I use an AI-native engineering workflow to move quickly from idea to implementation, while architecture, integration, validation, debugging and operational decisions remain under my direction.
SELECTED TECHNICAL SYSTEMS
Multi-Engine Quantitative Trading & Market-Intelligence Platform
Designed and developed a large Python/MetaTrader architecture that separated observation, context, planning, timing, execution authority, management, replay, research and learning into specialized components instead of one monolithic trading script.
- Market intelligence. Session, liquidity, structure, bias, market-regime, QT-context, hypothesis, situation-memory and confidence-oriented engines.
- Execution architecture. Multi-symbol planning, active-plan synchronization, symbol registries, risk/veto layers, MetaTrader 5 integration and MQL5 execution boundaries.
- Research & replay. Historical and counterfactual replay, MFE/MAE path analysis, outcome intelligence, shadow/ghost research and behavioral diagnostics.
- State & validation. Structured JSON/JSONL state, persistent journals, diagnostic traces, geometry/context guards and evidence-driven promotion of automation.
XAUUSD Trading Research, Replay & Outcome Architecture
Built a separate XAUUSD-focused system around deterministic opportunities, lifecycle tracking, replay and bounded execution, with Python owning intelligence/state and MQL5 owning execution management.
- Opportunity lifecycle. Immutable approved opportunities, deterministic identifiers, append-only JSONL persistence, idempotency and conflict detection.
- Outcome engine. Restart-safe tracking, explicit terminal states, configurable expiry, MFE/MAE measurement and BUY/SELL symmetry validation.
- Replay & safety. Chronological replay, persistent resume, fail-closed handling of ambiguous intrabar evidence and explicit execution boundaries.
- Runtime direction. Multi-account architecture, account discovery/registry work and read-only observability while live order-send remained evidence-gated.
AI Agent Orchestration & Desktop Execution Platform
Forked and reverse-engineered an existing StarNet baseline, then substantially re-engineered it into ARKNET: an independent desktop orchestration system with changed product behavior, integrations, controls, runtime surfaces and release safeguards.
- Orchestration. Agent/subagent lifecycle, provider/model routing, execution profiles, concurrency controls, fallbacks and persistent run state.
- Tools & connectors. MCP client/server infrastructure, OAuth, schema handling, HTTP/stdio transports, skills and controlled browser/computer/code/filesystem/shell tool surfaces.
- Governance. Capability permissions, consent boundaries, workspace safety, budgets, cost controls, checkpoints, journals and recovery paths.
- Desktop & validation. Tauri/Rust shell, JavaScript/TypeScript application layers, automated tests, evaluation gates and build/release workflows.
Linux/VPS Verification & Deployment Control Plane
Built and operate the production system behind this portfolio, exposing sanitized machine-readable proof from a real Linux VPS instead of decorative telemetry.
- Runtime. Nginx delivery, systemd services/timers, Python collectors and deterministic public JSON snapshots.
- Release control. Immutable releases, atomic symlink switches, rollback and append-only deployment lineage.
- Integrity. Tamper-evident event/evidence chains, release validation and fail-closed truth boundaries between backend state and browser presentation.
- Public surface. Live status contracts, read-only control-plane inspection and explicit separation between implemented capability and authority.
SELECTED EXPERIENCE
Systems Builder
Project-led software and systems engineering using an AI-native workflow, with requirements, architecture, integration, testing, debugging and final operational decisions directed by me.
Earlier Professional Background
Film-production operations: freelance lighting, followed by camera-department stock and equipment operations.
ENGINEERING CAPABILITIES
State, Boundaries & Lifecycle
Modular responsibilities, explicit authority boundaries, persistent state, append-only records, deterministic identifiers, idempotency, restart reconstruction and fail-closed behavior.
Agents, Models & Tools
Provider/model routing, agents and subagents, MCP/connectors, permissions, budgets, tool execution, OAuth integrations, recovery and bounded autonomy.
Research, Replay & Execution
Market-context pipelines, MetaTrader integration, multi-symbol/multi-account architecture, trade-lifecycle state, replay, outcomes, MFE/MAE and diagnostic research tooling.
Windows + Linux Operations
Hands-on Windows Server/VPS and Linux VPS work, including database-backed Shaiya game-server environments, Nginx, systemd, service troubleshooting, deployment and rollback workflows.