Research Tools

Open-source tools for biomedical research, data pipelines, AI workflows, and multi-agent systems. Built to solve real problems, shared to enable others.

Total Tools

8

Complete

3

In Progress

5

Open Source

100%

StreamSense

✓ Complete

Real-time biomedical signal processing for yoga mindfulness research

A comprehensive framework for capturing, processing, and analyzing physiological signals during yoga and mindfulness practices. StreamSense integrates EEG, ECG, respiration, and other biomedical sensors to quantify the neurophysiological effects of contemplative practices.

Developed

2022-2023

Technologies

PythonSignal ProcessingMNE-PythonPandasNumPyMatplotlib

Key Features

  • ▸Multi-modal sensor integration (EEG, ECG, respiration, GSR)
  • ▸Real-time signal quality assessment and artifact detection
  • ▸Automated feature extraction (heart rate variability, spectral power, coherence)
  • ▸Session-based analysis with pre/during/post comparisons
  • ▸Statistical reporting with visualizations (time-series, topographic maps, spectrograms)
  • ▸Export to standard formats (BIDS, CSV, HDF5) for downstream analysis
📦GitHub

ProSense

✓ Complete

Complementary analysis toolkit for biomedical research protocols

ProSense extends StreamSense with advanced protocol management and complementary analysis methods. It provides tools for designing experimental protocols, managing participant data, and conducting complex statistical analyses including time-frequency analysis, connectivity metrics, and machine learning-based state classification.

Developed

2023

Technologies

PythonScikit-learnSciPyStatsmodelsSeabornJupyter

Key Features

  • ▸Protocol design and management (session templates, timing control)
  • ▸Participant database with demographic tracking and consent management
  • ▸Advanced connectivity analysis (coherence, phase-locking value, transfer entropy)
  • ▸Machine learning pipelines for state classification (meditation vs. baseline)
  • ▸Statistical testing suite (parametric, non-parametric, multiple comparison correction)
  • ▸Automated report generation with publication-ready figures
📦GitHub

Cognitive Load Induction

✓ Complete

PsychoPy-based experimental paradigm for stress and cognitive load studies

A standardized experimental framework built with PsychoPy for inducing and measuring cognitive load and stress responses. Includes multiple validated tasks (N-back, Stroop, mental arithmetic) with synchronized physiological recording triggers, enabling investigation of how cognitive demand affects neural and cardiovascular function.

Developed

2022

Technologies

PythonPsychoPyLSL (Lab Streaming Layer)OpenCVPygame

Key Features

  • ▸Battery of cognitive tasks (N-back, Stroop, Serial 7s, Dual-task paradigms)
  • ▸Difficulty adaptation based on performance (staircase procedures)
  • ▸LSL integration for synchronized biomedical signal triggers
  • ▸Real-time performance metrics and adaptive difficulty
  • ▸Questionnaire modules (NASA-TLX, stress scales) integrated with task flow
  • ▸Data export with trial-level timing and accuracy logs
📦GitHub

Automated Data Pipeline

⏱ In Progress

End-to-end data scraping, processing, and dashboard visualization

A modular data pipeline for collecting, cleaning, transforming, and visualizing data from multiple sources. Originally built for research data aggregation, it generalizes to any data workflow requiring automated ETL (Extract, Transform, Load) processes with real-time monitoring dashboards.

Developed

2023-2024

Technologies

PythonApache AirflowPandasSQLAlchemyPostgreSQLPlotly DashDocker

Key Features

  • ▸Modular scraper architecture (web scraping, API calls, file monitoring)
  • ▸Data validation and quality checks with alerting
  • ▸Transformation pipelines (normalization, aggregation, feature engineering)
  • ▸Database integration with versioned schema migrations
  • ▸Interactive dashboards (Plotly Dash) with real-time updates
  • ▸Airflow orchestration for scheduled and event-triggered workflows
  • ▸Dockerized deployment for reproducibility across environments
📦GitHub

Material Ingestion Pipeline

⏱ In Progress

Knowledge graph construction from educational materials with Gen AI and MAS

An intelligent pipeline for processing educational documents (PDFs, slides, videos) into structured knowledge graphs. Uses NLP for entity extraction, Gen AI for semantic enrichment, and Multi-Agent Systems (MAS) for coordinated processing across document types. Designed to automate the creation of interconnected learning resources.

Developed

2024-2025

Technologies

PythonLangChainOpenAI GPT-4spaCyNeo4jVector Databases (Pinecone)Multi-Agent Systems

Key Features

  • ▸Multi-format document parsing (PDF, PPTX, DOCX, video transcripts)
  • ▸NLP-based entity and relationship extraction
  • ▸Gen AI-powered semantic enrichment (concept explanations, analogies)
  • ▸Knowledge graph construction in Neo4j with relationship inference
  • ▸Vector embeddings for semantic search and retrieval
  • ▸Multi-Agent System for task coordination (extraction → enrichment → graph construction)
  • ▸Query interface for exploring interconnected concepts
📦GitHub

Content Research Pipeline

⏱ In Progress

AI-powered content aggregation with vector search and multi-agent coordination

A research automation system that uses Gen AI and Multi-Agent Systems to aggregate, analyze, and synthesize information from diverse sources. Agents specialize in different research tasks (web search, paper analysis, fact-checking) and collaborate to produce comprehensive research summaries with source attribution.

Developed

2024-2025

Technologies

PythonLangChainAnthropic ClaudeChromaDBBeautifulSoupSeleniumMulti-Agent Systems

Key Features

  • ▸Multi-agent research team (Searcher, Analyzer, Fact-Checker, Synthesizer)
  • ▸Web scraping with dynamic content handling (JavaScript rendering)
  • ▸Academic paper parsing (arXiv, PubMed, Google Scholar)
  • ▸Vector database storage for semantic retrieval
  • ▸Source attribution and citation management
  • ▸Collaborative summarization with cross-agent verification
  • ▸Markdown report generation with structured sections and references
📦GitHub

Media Generation Pipeline

⏱ In Progress

Automated video content creation using Gen AI and multi-agent orchestration

A creative automation pipeline that generates video content from text prompts. Multi-Agent Systems coordinate scriptwriting, voiceover generation, visual asset creation, and video editing. Built for educational content creation, but generalizable to marketing, social media, and storytelling applications.

Developed

2024-2025

Technologies

PythonOpenAI GPT-4ElevenLabs TTSStable DiffusionFFmpegMoviePyMulti-Agent Systems

Key Features

  • ▸Multi-agent creative team (Scriptwriter, Voice Artist, Visual Designer, Editor)
  • ▸Script generation with scene breakdown and timing
  • ▸Text-to-speech voiceover with emotion and pacing control
  • ▸Visual asset generation (images, animations) from scene descriptions
  • ▸Automated video editing (scene assembly, transitions, subtitles)
  • ▸Style templates for different content types (explainer, tutorial, story)
  • ▸Iterative refinement based on quality assessment agents
📦GitHub

AI Workflow Portal

⏱ In Progress

Codebase analysis and multi-agent system orchestration with MCP integration

A comprehensive platform for analyzing codebases and orchestrating Multi-Agent Systems (MAS) through the Model Context Protocol (MCP). Provides tools for code documentation, architecture visualization, automated refactoring suggestions, and coordinated agent workflows for complex development tasks.

Developed

2024-2025

Technologies

PythonTypeScriptLangChainAST ParsingNeo4jModel Context Protocol (MCP)Multi-Agent Systems

Key Features

  • ▸Codebase ingestion and AST-based analysis
  • ▸Automated documentation generation (docstrings, architecture diagrams)
  • ▸Dependency graph visualization and impact analysis
  • ▸Multi-Agent System for coordinated development tasks
  • ▸MCP integration for standardized agent communication
  • ▸Refactoring suggestions with automated PR generation
  • ▸Code quality metrics and technical debt tracking
  • ▸Interactive chat interface for codebase Q&A
📦GitHub

Why Build Open-Source Research Tools?

Research tools should be accessible, reproducible, and improvable by the community. Every tool here started as a solution to a real research problem—then evolved into something others could use and extend.

🔬

Reproducibility

Open code means open science. Others can replicate methods, verify results, and build confidence in findings.

🌍

Accessibility

Free tools democratize research. Labs with limited budgets can still do cutting-edge work.

🚀

Collaboration

Community contributions improve tools faster than any single developer could. Bug fixes, feature requests, and extensions emerge organically.

Want to Collaborate?

These tools are actively maintained and evolving. Whether you want to contribute code, report bugs, request features, or discuss research applications—your input is welcome.

Get in Touch →View GitHub Profile