a hardcore Logistician, with lots of love for Claude, Data and Analytics.

industrialEngineering background with an MSc in Logistics & Supply Chain Management

self taught on data exploration & analysis, as well as data visualization.

bridging data & logistics to define business decisions and improve processes.

main hobby: options chains analysis and greeks exposure interpretation

it all started with Excel and PowerQuery&Pivot!

Engineered a production-grade real-time options data pipeline leveraging Python's asyncio and websockets for concurrent I/O operations, processing 50+ messages/sec from Interactive Brokers WebSocket API with intelligent contract filtering (3.5% ATM strike range, 3 nearest expirations).

Architected a multi-worker message processing system (4 parallel workers with asyncio.Queue) implementing priority-based message routing, LRU caching (@lru_cache with 10K entry capacity) for field parsing optimization, and connection pooling (aiohttp sessions) for API efficiency.

Data pipeline outputs structured time-series data to Polars DataFrames with explicit schema enforcement (17 typed columns: Greeks, IV, volume, OI, bid/ask), persisted incrementally to Parquet format (pyarrow) with 15-minute file rotation boundaries.

Implemented comprehensive data quality controls including regex-compiled field cleaners, NaN handling, and automated schema validation. Performance monitoring includes real-time metrics tracking (msg/sec throughput, queue saturation, cache hit rates >90%, dropped message counts) with 30-second interval logging for operational observability.

Engineered an automated cryptocurrency trading system on ByBit leveraging Python with Polars for high-performance data transformations and Hidden Markov Models (HMM) via hmmlearn for probabilistic regime detection.

Designed a multi-stage data pipeline integrating REST API data ingestion, real-time feature engineering (LDPM indicators, ATR, EMA), and Discord webhook-based approval workflows with asynchronous I/O.

Implemented enterprise-grade risk controls including position sizing algorithms, trailing stop management, and exchange state synchronization, while maintaining comprehensive logging and error handling for production resilience.

Ingestion Layer: ByBit REST API → JSON parsing → Schema validation

Transformation Layer: Polars aggregations → Feature engineering → Statistical modeling

Decision Layer: HMM inference → Signal scoring → Rule-based filtering

Approval Layer: Discord webhook → Human confirmation → Event callback

Execution Layer: Order placement → State sync → Position reconciliation

Monitoring Layer: Logging, error handling, performance tracking

• Real-Time Analytics Pipeline

Built automated ETL workflows processing parquet-based market data using Polars lazy evaluation, implementing complex window functions and aggregations to compute 30+ derived metrics for real-time trading analytics

• Production Streamlit Dashboard

Developed scalable BI application with modular data pipelines, auto-refresh capabilities, and interactive Plotly visualizations supporting multiple data sources and configurable lookback periods