RautioProjects.
Let's talk

02 / Projects

Selected data, infrastructure and AI projects.

A collection of professional and self-directed projects. Use the shared labels to browse related work across this project catalogue and the blog.

Browse by tag

Select one or more tags to show related projects. All tags stay visible while browsing.

All projects

Problem, solution, technologies, role and lessons learned in one view.

  1. 06

    Home Assistant Data ETL Pipeline to Cloud PostgreSQL

    A Python ETL pipeline that transfers numeric Home Assistant sensor history from a self-hosted PostgreSQL database to a cloud warehouse.

    Problem
    Needed to keep Home Assistant local while making selected historical sensor data available for cloud-based reporting and analysis.
    Solution
    Built a scheduled Python pipeline that extracts new numeric sensor states incrementally, transfers them through an SSH tunnel, and loads a Bronze/Silver/Gold PostgreSQL warehouse with run and step monitoring.
    Technologies
    Python · PostgreSQL · SSH · psycopg · Medallion · Bronze/Silver/Gold · cron
    Role
    Data pipeline designer & implementer
    What I learned
    Incremental boundaries, transaction handling, monitoring, and explicit cleanup make even a small ETL job reliable and recoverable.

  2. 05

    Finnish Postal Area Sales Analysis and Mapping

    A Jupyter Notebook workflow for combining Finnish geographic data with sales data and interactive maps.

    Problem
    Needed to combine geographic postal area data with municipality information and visualize sales across Finnish regions.
    Solution
    Built a Jupyter Notebook that retrieves geographic and municipality data from Statistics Finland, generates and aggregates sales data, exports results to Excel, and creates interactive Folium maps by postal area and municipality.
    Technologies
    Python · Jupyter · Pandas · GeoPandas · Folium · Statistics Finland APIs
    Role
    Data analyst & visualization developer
    What I learned
    Combining geographic data with business metrics makes regional patterns easier to explore, communicate, and validate.
    Read more:
    View GitHub ↗

  3. 07

    Data Quality Analysis Tool

    A Python-based tool for profiling CSV and Excel datasets, surfacing quality metrics, distributions, correlations and categorical patterns.

    Problem
    Needed a repeatable way to understand dataset structure and identify practical data quality issues before analysis or reporting.
    Solution
    Built a profiling workflow with pandas, NumPy, Seaborn and Matplotlib, supported by an AI-assisted analysis process that turns findings into strategic and operational data quality insights.
    Technologies
    Python · Pandas · NumPy · Seaborn · Matplotlib · Data profiling
    Role
    Data quality analysis designer & implementer
    What I learned
    Combining automated profiling with critical human review makes data quality findings more useful for operational decisions.

  4. 01

    Local Git Server on Synology NAS

    A local Git server with SSH access and automated backups for secure code and configuration management.

    Problem
    Need to manage code and configuration in a secure, local environment.
    Solution
    Set up a local Git server on a Synology NAS with SSH access and automated backups.
    Technologies
    Synology NAS · Git · SSH · Bash scripting
    Role
    Solution designer & implementer
    What I learned
    Local solutions can be powerful when security and control are priorities.

  5. 02

    Self-Hosted Prototyping and Demo Environment

    A production-inspired homelab for prototyping, service hosting, and technical demonstrations.

    Problem
    Needed a safe environment to design, test, and showcase infrastructure and software solutions.
    Solution
    Built a production-inspired homelab platform on an HP Mini PC for prototyping, service hosting, and technical demonstrations.
    Technologies
    HP Mini PC · Linux(Debian) · Containers · Automation · Networking · Monitoring · Documentation · Security · VPN
    Role
    Solution designer & implementer
    What I learned
    Testing architectures locally enabled faster cloud deployments and more predictable implementation outcomes.

  6. 03

    Cost-Optimized Multi-Agent AI System

    A hierarchical agent architecture that combines coordinator models, delegated sub-agents, and Telegram interaction.

    Problem
    Running advanced AI workflows with large models alone was expensive and inefficient.
    Solution
    Designed a hierarchical agent architecture with Hermes, using a powerful coordinator model and delegated sub-agents for token-intensive workloads. Added Telegram integration for real-world interaction.
    Technologies
    Hermes · OpenClaw · APIs · Telegram · Multi-Agent Systems · Linux
    Role
    AI solution architect & implementer
    What I learned
    Combining orchestration, delegation, and parallel execution greatly improves both cost efficiency and overall system performance.

  7. 04

    From Homelab Prototype to Cloud Deployment

    A secured cloud-hosted platform for running services, managing traffic, and supporting real-world workloads.

    Problem
    Needed to transition validated concepts from a local testing environment into a production-ready platform.
    Solution
    Built and secured a cloud-hosted infrastructure for running services, managing traffic, and supporting real-world workloads.
    Technologies
    Cloud Platform · Linux · Containers · DNS · Security
    Role
    Infrastructure architect & operator
    What I learned
    A well-tested foundation enables faster and more predictable production deployments.