Built
Proof, not promises.
A studio is only as good as what it ships. Here is what I have built and run, from products to AI agents. The advisory work draws on all of it.
Product
ShippedSkylark Client IQ
- Problem
- Sales leaders lack a single view of rep performance and which client relationships are quietly going at risk.
- What I built
- A Salesforce-integrated sales performance and client-risk platform, with custom business logic and a risk-scoring model, delivered as three role-based dashboards.
- Outcome
- Designed and shipped end to end for a client.
- Salesforce (SFDC) integration as the system of record
- Three dashboards by persona: sales rep, manager, and executive team
- Custom business logic modeled to the client's sales motion
- Client risk-scoring model that surfaces at-risk accounts early
The data-integrity and risk logic runs on the same governance backbone I open-sourced as the Data Steward Agent ↗.
Agent
Open sourceData Steward Agent
- Problem
- Teams sit on messy data and lack a safe, repeatable way to let an agent clean and govern it.
- What I built
- An open-source agent that profiles, cleans, and documents datasets with a human in the loop.
- Outcome
- Released as open source for other builders to run and adapt.
- Three blocking gates run inside the delivery pipeline, not a retrospective review
- Data integrity gate: quality, consistency, documentation, and business context
- Privacy and security gate: PII, GDPR and CCPA, and deletion protocols
- Change management gate: schema governance with validated rollback, public on GitHub with diagrams and templates
Product
ShippedGoals App
- Problem
- Personal goals scatter across notes and apps, so progress is hard to see and easy to drop.
- What I built
- A focused goal-tracking app, designed and built end to end, live on its own subdomain.
- Outcome
- Shipped and running in production.
- Twelve goals across four categories, three per category, by deliberate constraint
- Fully client-side: no accounts, no database, no backend to maintain
- Built on a pre-decided Skylark 118 design system for a 3-hour intent-to-live ship
- Export, share, pin, or set as a phone wallpaper to keep goals visible
Agent
ArchivedSignal Stream
- Problem
- Marketing teams miss buyer signals because monitoring is manual and easy to let slip.
- What I built
- A market-intelligence agent that monitored public conversation, scored signal, and delivered digests.
- Outcome
- Ran a live intelligence feed into Slack for three months, then was retired with the engagement it served. The architecture is the reference for Ronin.
- Ran unattended on a GitHub Actions pipeline from March to May 2026
- Collected roughly 147 community posts per run through a residential proxy
- Scored relevance with Claude Haiku layered on keyword scoring
- Delivered daily and weekly signal digests straight into Slack
About the work
- Does Linda actually build, or just advise?
- Both. Everything on this page was designed and built hands-on, from a live goals app to open-source agents and production market-intelligence systems. The advisory work is grounded in the same building, not borrowed from someone else's case studies.
- Can I use any of these?
- Some, yes. The Goals App is live and the Data Steward Agent is open source. Others are products and client systems at various stages. If one is relevant to your situation, mention it on a discovery call.
Want something like this built?
Whether it is an agent, a dashboard, or an advisory engagement to put agents to work across your team, the next step is the same: a short conversation about what would move the needle.
Book a discovery call