Data & Signal Processing Development Services

Building Scalable Python Platforms with TDD Frameworks and Long-Term Support

About

I am Noam Shabtai, with a PhD from Ben-Gurion University (Prof. Boaz Rafaely) and postdoctoral research at RWTH Aachen University (Prof. Michael Vorlaender), specializing in array signal processing, real-time systems, and spectral analysis.

This academic foundation in mathematical signal processing - including multi-channel sensor arrays, frequency-domain transformations, and real-time algorithm optimization - applies across domains: audio, medical imaging, radar/lidar, communications, and industrial sensing.

I provide professional Python-based development and testing services for signal processing applications. My expertise lies in building scalable platforms with comprehensive Test-Driven Development (TDD) frameworks and providing long-term support that ensures your systems remain maintainable and reliable.

I help organizations transform their signal processing algorithms into production-ready, scalable platforms with robust testing infrastructure and modern development practices.

Services I Offer

Custom Python Infrastructure: Building scalable platforms tailored for signal processing applications
TDD Framework Development: Establishing comprehensive testing frameworks that ensure code quality and accelerate development cycles
Long-Term Platform Support: Providing ongoing maintenance, updates, and knowledge transfer to your team
Real-Time Processing Systems: Production-ready frameworks with live streaming and data processing capabilities
Algorithm Implementation: Transforming research algorithms into production-ready, tested code
Modular Architecture Design: Scalable systems with clean separation of concerns and reusable components

Portfolio Demonstration

Example Project: Real-Time Signal Processing Framework with Spatial Audio Application

To demonstrate the quality and depth of services I provide, I have built a complete, production-ready signal processing infrastructure. This framework provides general-purpose components for any signal processing application, with spatial audio implemented as an example use case showcasing the platform’s capabilities.

The infrastructure demonstrates how modular, scalable platforms enable rapid development of signal processing applications while maintaining high code quality and testability.

Core Infrastructure Components

System Architecture: Base system class providing module connection and execution framework - applicable to any signal processing pipeline
Buffer Management: Sophisticated input/output buffer handling for streaming data in real-time applications
Activator Pattern: Abstract base class with both file-based and real-time callback implementations for flexible deployment across different use cases
STFT Processing Module: Modular Short-Time Fourier Transform with separate Analysis and Synthesis classes, perfect reconstruction, and configurable overlap ratios (2x, 4x, custom)

Example Application: Spatial Audio

Built on top of the general infrastructure to demonstrate real-world usage:

Spatial Audio System: HRTF-based binaural rendering for immersive 3D audio with quaternion-based head orientation tracking
Real-Time GUI Demo: Tkinter-based interface with live azimuth/elevation control and per-channel gain management

Try it yourself (requires Python 3.12+, uv, and any standard headphones):

git clone https://github.com/noamshabtai/signal-processing.git
./signal-processing/spatial-audio-demo/run_demo.sh

This example shows how the infrastructure enables rapid development of complex signal processing applications while maintaining code quality and testability.

Development Quality

Comprehensive Testing: 100+ tests across all modules using pytest with YAML-based parametrization
CI/CD Integration: GitHub Actions with automated testing on all pull requests, branch protection requiring passing tests
Modern Development Practices: Pre-commit hooks, type hints throughout, Python 3.12+, uv package management
Modular Design: Clean separation of concerns with multiple interdependent packages in a monorepo structure
Production-Ready: Real-time processing capabilities, perfect reconstruction guarantees, configurable parameters

Platform Capabilities

General Infrastructure:

Frequency-domain processing with configurable STFT pipeline
Real-time streaming with callback-based architecture
File-based processing for offline analysis
Modular design allowing easy addition of new processing modules

Example Application (Spatial Audio):

Multiple virtual sound sources at configurable 3D positions (azimuth/elevation)
Input: CH channels (mono sources) → Output: Binauralized stereo
Real-time streaming with PyAudio integration
File I/O supporting .wav and .bin formats

Source Code & Documentation: github.com/noamshabtai/signal-processing

The complete framework is open-source, demonstrating transparency and code quality. Technical documentation and architecture notes are included in the repository.

Data Processing Framework

Example Project: Financial Data Analysis with ML Predictions

Building on the signal-processing infrastructure, this project demonstrates how the same architectural patterns apply to data processing applications. The finance demo shows cross-domain reuse of the activator pattern for polling-based data pipelines.

Components

Data Fetcher: Stock data retrieval using yfinance with historical and real-time capabilities
Feature Extraction: Signal processing techniques applied to financial time series - trend extraction and FFT-based frequency analysis
Model: PyTorch LSTM for sequence prediction
Finance Demo: Inherits from signal-processing’s base_demo activator, demonstrating code reuse across domains
Stock Analyzer CLI: Command-line interface for training and prediction

Architecture Highlights

Cross-Repository Dependencies: data-processing depends on signal-processing subpackages (activator, system, buffer)
Shared Test Infrastructure: Uses the same YAML-based parametrize-tests framework
Inherited Patterns: Finance demo activator inherits running, process_frame(), stop() from signal-processing’s base_demo

Source Code: github.com/noamshabtai/data-processing

Get in Touch

If you’re interested in building scalable signal processing platforms for your organization, I welcome the opportunity to discuss how my services can support your needs.

Whether you need custom infrastructure development, TDD frameworks, or long-term platform support, I can help transform your signal processing requirements into production-ready solutions.

Contact Information

Phone: +972-58-448-8767
Email: shabtai.noam@gmail.com
LinkedIn: linkedin.com/in/noam-shabtai-80836717
GitHub: github.com/noamshabtai