On top of Mauna Kea

Tools I use and gadgets I like.

I've used a number of tools in work and everyday life and compiled a list of recommendations. These back the skills I've developed and the experience I've gained across my work and projects.

Product Management

  • GitHub Projects

    This has been my preferred product management tool since it shipped. Boards span multiple repositories, custom fields go beyond the default issue set, and linking pull requests to issues keeps Product and Engineering working off the same board.

    docs.github.com

  • Jira

    This is my tool of choice for larger, multi-stakeholder products. You can define epics and link issues into a proper roadmap, and its backlog and sprint management scale past a single team without much friction.

    atlassian.com/software/jira

  • Asana

    I reach for Asana for cross-team task tracking when a project doesn't need Jira's overhead, which is more often than people expect.

    asana.com

User Research & UX

  • Figma

    Figma is my go-to for wireframes and managing a design system. Its prototyping mode turns static screens into something a stakeholder can actually click through.

    figma.com

  • Dovetail

    I've used Dovetail for years to turn user research interviews into themes, personas, and journey maps. It isn't cheap, but I haven't found anything else that compiles research into decisions this well.

    dovetail.com

  • Miro

    Miro is where the workshops and stakeholder mapping happen before any of it turns into a requirement someone builds against.

    miro.com

Data Architecture & Engineering

  • Airflow

    Apache Airflow is my default for orchestrating integrations between systems. It uses DAGs to make dependencies explicit, so scheduling stays simple even once you're running more than a couple of pipelines. It's open source, so there's no license standing between you and another pipeline.

    airflow.apache.org

  • Microsoft Power Platform

    Plenty of NGOs already run on Microsoft 365, and the Power Platform lets them build without hiring developers: PowerApps for internal tools, Power Automate for integrations, Power Query for reshaping data. The trade-off is the same one that applies to any low-code platform. It shifts cost from custom development to licensing and ties you into Microsoft's ecosystem, which is a fair deal for a lean IT team, but worth naming out loud before committing.

    powerplatform.microsoft.com

  • OpenFN

    An open-source integration platform built specifically for the humanitarian and development sector, with adapters for case management and M&E systems most general-purpose ETL tools have never heard of.

    openfn.org

  • Swagger

    For API-first design, Swagger is still the standard for documenting and validating an API contract before anyone writes an integration against it.

    swagger.io

Databases & Enterprise Systems

  • PostgreSQL & MySQL

    For structured, relational data, PostgreSQL is my default: stronger typing, a solid extension ecosystem, and no real downside once you're past a basic install. MySQL still shows up plenty, usually because it's already what a host or CMS assumes, and it's a fine choice when that's the case.

    postgresql.org

  • Airtable

    Airtable bridges the gap between a spreadsheet and a database. Table relationships, validation, and conditional formatting come built in, without standing up a real backend.

    airtable.com

  • Google Sheets

    I reach for Sheets over Excel by default. It has the same functions, plus real-time collaboration without emailing a file back and forth.

    google.com/sheets

  • Enterprise & Low-Code Platforms

    I've worked across MS Dynamics 365, Odoo, Salesforce, and Unit4, enough to know each one speaks its own proprietary configuration language. The right pick usually matches whatever an organization already runs on, not a personal favorite. It's worth flagging either way that these low-code platforms shift cost from custom development to licensing, which is a fair trade for a lean team, but it is vendor lock-in: budget for it, and expect migrating off one later to be expensive.

    salesforce.com

  • M&E Platforms

    ActivityInfo and DHIS2 are open source, so you avoid licensing costs but take on hosting and maintenance yourself. DevResults and RedRose are proprietary and turnkey, which suits a team that doesn't want to run infrastructure. TolaData sits in between the two, and it's the one I actually helped build.

    activityinfo.org

Analytics & Business Intelligence

  • Open Source BI: Superset or Metabase

    Superset and Metabase are my defaults when a client wants to avoid another subscription line item. Superset scales further for a team willing to self-host and tune it, while Metabase gets a smaller team to a usable dashboard faster.

    superset.apache.org

  • Power BI

    Power BI is the natural pick once an organization already lives inside the Microsoft ecosystem, and it's a genuinely powerful tool. I've also seen it used just as often to build dashboards that are complex rather than clear, and its cost and vendor lock-in are worth weighing before you commit.

    powerbi.microsoft.com

  • Google Looker Studio

    The Google-stack equivalent, and the obvious choice once an organization is already living in Google Workspace or BigQuery. It's also the fastest path I know from a data source to an interactive dashboard when I need something client-facing on short notice.

    lookerstudio.google.com

  • Elastic

    I use Elastic specifically for event and log analytics, not general BI. Indexing application and infrastructure logs means a spike in errors is a query away, not a grep through files on a server.

    elastic.co

  • Tableau

    Tableau doesn't really belong to a stack the way the others do. I use it when a client already has it and knows how to drive it, since standing it up from scratch rarely pencils out against the license cost.

    tableau.com

  • Jupyter Notebooks

    This is my interface for most Python data work. Pandas and scikit-learn handle the wrangling and modeling, and the notebook format doubles as a written narrative around the analysis.

    jupyter.org

  • R

    For statistically intensive analysis, I prefer R and its package ecosystem. RMarkdown carries the narrative, and ggplot2 handles the visualization.

    r-project.org

Software Development & Cloud

  • VS Code

    VS Code is my default editor. An integrated terminal means package installs and debugging never have to leave the window.

    code.visualstudio.com

  • Frontend

    I write TypeScript across React and Vue for single-page applications, styled with Tailwind or SCSS depending on the project.

    react.dev

  • Tailwind CSS

    Utility classes and thorough documentation make this the fastest way I know to get an interface looking right without fighting a stylesheet.

    tailwindcss.com

  • Backend

    I use Node and Next.js for smaller apps, and Python with Django and DRF for anything data-heavy, where its library ecosystem really earns its keep.

    djangoproject.com

Monitoring and Evaluation

  • Open Data Kit (ODK)

    XLSForm authoring and ODK Central have carried most of the nonprofits I've worked with through the move to mobile data collection. Kobo, CommCare, and SurveyCTO fill in depending on the deployment.

    getodk.org

GIS & Spatial Data

  • QGIS

    QGIS is my default for spatial analysis. It's open source, so there's no license standing between an idea and a map.

    qgis.org

  • ArcGIS

    ArcGIS shows up when a client's GIS team already has an Esri license and workflows built around it. Rebuilding that in QGIS usually isn't worth fighting for.

    arcgis.com

  • Mapbox

    My choice for maps that ship inside a product, for both the styling and the hosting, built on OpenStreetMap data underneath.

    mapbox.com

AI & Machine Learning

  • MCP & Agent Tooling

    I build agent tooling on the Model Context Protocol. That work spans prompting and evaluation, fine-tuning small models, and retrieval-augmented generation to ground answers in real data.

    modelcontextprotocol.io

  • Ollama

    I use Ollama for running models locally, whether the data can't leave the building or a workflow just needs to run without an API dependency.

    ollama.com

IT Service Management

  • FreshService

    My default for ITIL-aligned service management at a smaller organization: incident, problem, and change management, plus a service catalog, without enterprise pricing attached.

    freshservice.com

  • ServiceNow

    ServiceNow does the same job at real enterprise scale, but it comes with enterprise pricing and usually its own dedicated admin. Worth it once a service desk has outgrown FreshService, not before.

    servicenow.com

Documentation

  • Markdown

    Markdown is the format behind most of my documentation. GitHub repos, internal product docs, and knowledge bases all read it natively, emojis included 😅.

    docs.github.com

  • GitBook

    GitBook keeps Markdown under version control, with a clean table of contents and search that make it easy for non-technical teammates to contribute. It's free for open-source projects and nonprofits, and self-hostable if you need that.

    gitbook.com

Work station

Curious how these tools might fit your stack? Get in touch.