About
arus impact is a fractional AI/ML practice building data infrastructure for nonprofits and civil society organizations across press freedom, child safety, digital rights, cybersecurity, and open knowledge.
The Builder
Twenty years building startups. Y Combinator Techstars. Early engineer to technical co-founder, Series A through acquisition — the full arc, many times over. Enough reps that the pattern is muscle memory, not theory.
The pivot to impact work was pattern recognition. The same data pipelines, ML infrastructure, and integration middleware I'd been building for companies optimizing engagement metrics could transform organizations tracking ransomware attacks, investigating war crimes, or protecting children from predatory tech. The nonprofits doing the hardest work were running on spreadsheets and manual processes — not because better tools didn't exist, but because nobody was building them for orgs operating on mission-driven budgets.
So I started building them. Senior engineer, hands on keyboard, embedded inside your org — writing the data pipelines, training the models, deploying the infrastructure that lets you focus on your mission instead of fighting your tools.
The Network
Organizations Researched
Every org scraped, enriched, and scored across a 6-dimension AI/ML fit framework — tech gap, impact potential, team size, domain relevance, build viability, and growth trajectory.
Projects Scoped
Each project grounded in deep research — actual tools the org uses, actual workflows that break, actual builds that would move the needle. Published transparently on this site.
Project Liberty Alliance
Warm introductions to leadership across 178 organizations via Frank McCourt's coalition connecting technologists with civil society. The network that turns research into relationships.
Domains
Bellingcat, RSF, American Journalism Project
Open-source intelligence pipelines, newsroom data infrastructure, investigative verification tools
Fairplay, Christchurch Call
Dark pattern detection, content classification, platform policy monitoring at scale
CyberPeace Institute, IST
Threat classification pipelines, policy recommendation tracking, ransomware incident analysis
Dream.org, Heartland Forward
Cross-program outcome tracking, economic impact analytics, community-level GIS mapping
Creative Commons, Metagov
License compliance infrastructure, governance data platforms, commons health analytics
Data-Pop Alliance, EngageMedia, CCDH
Early warning systems, multilingual NLP, internet shutdown detection, hate speech classification
Methodology
Scrape, enrich, and score organizations across a 6-dimension AI/ML fit framework. Quantitative analysis — not vibes — to find where infrastructure creates 10x value.
Every engagement starts as a public proposal on this site. Transparent scope, grounded in what the org actually uses today and what would actually move the needle.
Hands on keyboard, inside your org. Production data pipelines, ML models, integration middleware. Code that ships, not slide decks.
Tools built for one organization become available to all. Each engagement strengthens the ecosystem. The rising tide approach.
The Model
Senior engineering at a fraction of a full-time hire. Deliberately structured for nonprofits that need real infrastructure but can't justify a $200K engineering salary.
I write the code, build the pipelines, deploy the models. Inside your org, on your stack, shipping production infrastructure — not recommendations about what someone else should build.
Fractional engagements structured for nonprofit budgets. No SOWs, no enterprise pricing, no procurement theater. Scoped to your capacity so the budget conversation takes five minutes, not five months.
Tools built for one organization become available to all. The commons approach — each engagement strengthens the entire ecosystem, not just one org's infrastructure.
Every engagement starts as a conversation. If your organization is doing real impact work and your infrastructure is holding you back, I'd like to hear about it.