Projects

From bioinformatics to ecological AI systems

cat project_philosophy.txt

My work spans from open-source bioinformatics tools for the scientific community to cutting-edge AI systems that integrate ecological intelligence. Each project embodies my commitment to resilient, accessible technology that operates within planetary boundaries while enhancing human capabilities.

Current AI Edge Computing Projects

Active

BioDOG (Biological Database-Oriented Generation)

A framework that bridges protein language models with biological ontologies, enabling enhanced analysis of complex biological systems through context-aware AI.

  • Interfaces between protein language models and established biological ontologies
  • Creates structured representations of complex biological knowledge
  • Enables cross-modal reasoning between text and biological data
  • Designed for resource-efficient deployment in research environments
Python PyTorch ESM-2/ProtT5 Knowledge Graphs Gene Ontology
Started: February 2024 Status: Development
Active

Mahakala

A lightweight, edge-deployable AI system designed for resource-constrained environments with ecological awareness built into its architecture.

  • Optimized for deployment on low-power devices and intermittent connectivity
  • Implements novel resource-aware computing algorithms
  • Features built-in privacy protections and local-first operation
  • Adapts computational load based on available renewable energy
NixOS GGUF Models Rust TinyML ARM Compute
Started: March 2024 Status: Prototype
Active

Localized LLM Deployment Framework

Creating accessible infrastructure for community-oriented AI technology that respects privacy and operates independently of centralized cloud systems.

  • Reproducible deployment system using declarative NixOS configurations
  • Community-first design prioritizing accessibility and transparency
  • Integration with local knowledge bases and context repositories
  • Optimized for consumer hardware with minimal resource requirements
NixOS llama.cpp Docker Python LangChain
Started: January 2024 Status: Beta Testing
Concept

EcoAgent

An autonomous agent framework designed to monitor and assist with ecological restoration projects, combining edge sensing with AI-driven analysis and decision support.

  • Distributed sensor network integration for environmental monitoring
  • AI-driven analysis of ecosystem health and recovery patterns
  • Decision support for restoration interventions and adaptive management
  • Low-impact deployment designed for minimal ecological disturbance
IoT Edge AI Time Series Analysis Environmental Sensors
Status: Conceptual Design

Scientific Bioinformatics Projects

Legacy

MarFERReT (Marine Functional Reference Eukaryotic Transcriptomes)

An open-source, version-controlled reference library of marine microbial eukaryote proteins, addressing the underrepresentation of marine organisms in existing genetic databases.

  • Comprehensive collection with tens of millions of validated reference proteins
  • Preserved taxonomic granularity from strain to species level
  • Flexible framework for building custom reference libraries
  • Rigorous cross-validation to identify quality and contamination issues
Python BLAST BioPython Docker NCBI Tools
Published: 2023 Status: Maintained
Legacy

North Pacific Eukaryotic Gene Catalog

A comprehensive catalog of metatranscriptome assemblies and annotations for marine microbial eukaryotes in the North Pacific Ocean, providing essential reference data for oceanographic research.

  • Contains over 200 million environmental gene sequences with functional annotations
  • Region-specific focus on the North Pacific ecosystem
  • Facilitates collaborative research in marine microbial ecology
  • Illuminates patterns of gene expression across time and space
Metatranscriptomics RNA-Seq Trinity KEGG R
Published: 2024 Status: Available
Legacy

Diel Transcriptional Cascades

Investigation of the daily (diel) rhythms of gene expression in open-ocean eukaryotic plankton communities, revealing synchronized biological processes driven by light cycles.

  • Mapped temporal organization of gene expression in marine microbial communities
  • Identified key regulatory elements controlling diel oscillations
  • Demonstrated niche differentiation through temporal specialization
  • Connected gene expression to metabolic activity in the ocean
Time Series Analysis RNA-Seq Statistical Modeling R DESeq2
Published: 2021 Status: Completed

grep "continuity" project_trajectory.txt

The trajectory from bioinformatics to AI Edge Computing reflects a natural evolution—from analyzing complex biological systems to designing technological systems that embody biological principles of resilience, efficiency, and ecological harmony.