Selected Systems

These projects are presented as systems and architecture case studies rather than commercial offerings. The focus is not on deliverables, but on system design, control mechanisms, and architectural decisions under real-world constraints.

Across these systems, the common thread is architectural thinking: managing complexity, isolating risk, and designing stable systems that operate reliably over time.

Some of these systems are active trading architectures. Others are large-scale infrastructure and data systems built in different contexts. Together, they represent a body of work centered around system architecture, risk control, and operational design.

TRADING SYSTEMS

Global Risk Engine

Distributed risk and capital allocation architecture for multi-strategy trading

The Global Risk Engine is a specialized control system designed to coordinate multiple trading strategies within a shared capital environment. The architecture is inspired by institutional risk systems used in hedge funds and asset managers such as State Street, adapted for private trading environments.

The system is built as a distributed microservices architecture, coordinating multiple trading engines and risk modules across independent services. It integrates with Interactive Brokers via TWS and API interfaces to manage execution, exposure, and capital allocation.

The architecture focuses on maximizing capital efficiency while controlling correlations between underlyings, strategies, and exposure types. Strategy isolation, exposure limits, and centralized kill-switch logic ensure that individual trading layers cannot create uncontrolled portfolio-level risk.

Risk budgets are dynamically allocated across trading engines. Each engine operates independently but within globally enforced constraints, allowing for parallel strategy execution while maintaining strict portfolio-level risk control.

The system is designed for live capital deployment and continuous operation under real-world market conditions. Monitoring, fail-safe mechanisms, and execution guards ensure stable behavior even during volatility spikes or infrastructure disruptions.

The result is a multi-layer trading infrastructure bringing hedge-fund-style risk architecture to private trading environments.

Layer 1 Core Yield Trading Engine

Fundamental-driven premium-selling architecture for undervalued equities

Layer 1 forms the core yield engine of the trading architecture. The system identifies undervalued, high-quality equities using automated fundamental analysis and structured screening logic.

The fundamental engine evaluates financial metrics including cash flow, EBITDA, earnings growth, and gross margins across rolling time windows of one, three, five, and ten years. These metrics are combined into a systematic scoring model designed to identify stable companies trading below intrinsic value.

Once candidates are identified, the trading layer systematically sells short puts on selected underlyings. Assignment leads to long-term equity positions, which are then managed through structured covered call workflows.

Position lifecycle management is fully integrated into the system architecture. This includes entry logic, assignment handling, covered call management, and exit workflows. The system is designed to operate deterministically, minimizing discretionary decisions.

Layer 1 is implemented as a system monolith optimized for reliability and deterministic execution. The engine operates under live capital constraints and is designed for long-term stability rather than short-term optimization.

The architecture prioritizes capital preservation, predictable yield generation, and operational stability. Layer 1 acts as the foundational income layer within the broader multi-strategy trading system.

Layer 2 Yield Optimization Trading Engine

Defined-risk credit spread architecture for yield optimization

Layer 2 extends the Layer 1 universe by introducing defined-risk credit spreads. This layer focuses on improving capital efficiency while maintaining bounded exposure and clearly defined risk limits.

The engine selects underlyings from the Layer 1 fundamental universe and applies structured spread strategies with predefined maximum loss and capital allocation constraints. Risk exposure is explicitly limited at the trade level.

Layer 2 operates with shorter time horizons and more frequent position turnover compared to Layer 1. The architecture is designed to complement the core yield layer while maintaining strict portfolio-level risk discipline.

The system is implemented as a deterministic monolith optimized for reliability and predictable execution. Like Layer 1, it integrates directly into the Global Risk Engine for capital allocation and exposure control.

Layer 2 functions as a yield optimization engine within the multi-strategy architecture, improving capital efficiency without introducing uncontrolled risk.

SPX 0DTE Engine

Intraday options execution architecture for profit maximization

The SPX 0DTE engine is designed for intraday options trading with short decision cycles and strict risk constraints. The system operates entirely within same-day expiration windows and avoids overnight exposure.

Execution logic is fully rule-based and operates within predefined time windows, volatility regimes, and market conditions. The engine monitors volatility, trend behavior, and time decay dynamics to determine valid execution opportunities.

Risk constraints are enforced both locally and through the Global Risk Engine. Position sizing, maximum loss thresholds, and exposure limits are dynamically enforced to prevent excessive risk concentration.

The SPX engine operates with high-frequency decision cycles and requires robust infrastructure, deterministic execution, and strict fail-safe mechanisms. The architecture prioritizes controlled execution over aggressive positioning.

This layer functions as a tactical profit engine within the broader architecture, complementing the longer-term yield layers while maintaining strict risk discipline.

Integration with the Global Risk Engine ensures that intraday exposure remains aligned with overall portfolio risk constraints.

DATA & INFRASTRUCTURE SYSTEMS

Crawling & Data Pipeline Architecture

Large-scale distributed crawling architecture for Financialbot AG

This system was designed as a multi-layer data acquisition architecture using distributed crawling infrastructure. The platform operated on bare-metal servers and processed large volumes of structured and unstructured data.

The architecture included distributed crawlers, ingestion pipelines, normalization layers, and Elasticsearch-based indexing systems. RabbitMQ was used for queue-based processing and distributed workload coordination.

The result was a scalable data platform capable of continuous data ingestion and structured retrieval.

Technology stack: Python, MySQL, Elasticsearch, RabbitMQ, Bare-metal infrastructure

Distributed Infrastructure Platform

Large-scale distributed infrastructure architecture for Living Internet / Yellowgrey

This project involved designing and operating a distributed infrastructure across a large number of nodes. The architecture included deployment automation, monitoring systems, and operational safeguards.

The infrastructure operated across distributed environments with orchestration and monitoring layers. The system focused on maintaining operational stability and visibility across distributed infrastructure.

Technology stack: Python, MySQL, Elasticsearch, distributed node infrastructure

High-Volume Data Processing System

Large-scale ingestion and indexing architecture for Hoosh / NOVASOL

This system focused on high-volume data ingestion and large-scale indexing pipelines. The architecture handled sustained processing load while maintaining consistent indexing and retrieval.

The design emphasized throughput, reliability, and structured access to large datasets. This created a scalable data processing environment.

Technology stack: Python, MySQL, Elasticsearch, distributed processing pipelines

AUTOMATION & BUSINESS SYSTEMS

Distributed Engineering Architecture

Remote engineering and delivery architecture for LI Ukraine

This architecture focused on distributed engineering teams operating across locations. The system defined delivery workflows, engineering standards, and coordination structures.

The goal was to maintain consistent engineering output across distributed teams. This resulted in a structured delivery architecture.

Pharmacy Digitalization Architecture

Operational workflow and communication architecture for Montanus Apotheke

This system focused on replacing manual operational workflows in a pharmacy environment. Previously fragmented communication and document handling were replaced with structured digital workflows.

The architecture centered on process design, validation logic, and integration between operational steps. This created a more predictable and scalable operational environment.

WHY THIS MATTERS

Architecture made visible through systems.

Most architecture work remains invisible. It sits behind execution, quietly shaping how systems behave under pressure.

These selected systems make that layer visible.

The point is not polished launches or product narratives. It is to show how architecture decisions translate into working systems: how constraints are handled, how complexity is reduced, and how control is maintained as systems evolve.

Each case reflects a different type of system: trading architecture, data infrastructure, distributed platforms, or operational workflows.

Together, they represent a consistent approach: designing systems that remain stable, understandable, and adaptable over time.