AI-native product builder · Systems that ship · Cleveland, OH

Designed. Built.
Delivered.

Ten years turning data into decisions at Cigna, Salesforce, and Deloitte. Now I design, engineer, and ship AI-native products end to end: a private AI workforce for industry, a front office for all 30 MLB clubs, rehab software for the training room, and a meal planner for one very hungry family.

See the work Download resume
№1Inner Pocket №2Keystone №3Rebound №4What Should We Make №5LUCI Explorer №6Good Vibes Club

Selected work · 2025–2026

An index of shipped work

Every row is live today

01 Inner Pocket 30,000+ regulation sections cited verbatim, in production for client №1 AI platform 02 Keystone A front office for all 30 MLB clubs, every model graded in public Decision platform 03 Rebound 27 real care protocols and a 100-exercise library, live at rebound.care Care operations 04 What Should We Make Family meals planned in 60 seconds, from 225+ curated recipes, free forever Consumer 05 LUCI Explorer A heliocentric map of 929 tokens, adopted by Sam Spratt for his site redesign Ecosystem viz 06 Good Vibes Club Analytics 14,346 ETH of market history, self-updating, every chart share-ready Market intel

01 · AI platform

Inner Pocket

A full staff, kept close: private AI specialists, designed to order for each business, answering with the regulation text open on the table

Each client's bench is drawn from a library of fourteen roles. The flagship is a Regulations Expert wired to a local library of 30,000+ federal and state regulation sections: every answer cites the rule, and clicking a citation opens the verbatim text. A public read-only showroom, staffed for a fictional metal-finishing shop, demos it all with real product answers.

The launch film: a full staff, kept close. Fictional demo client, real product answersFig. 01 of 06

02 · Decision platform

Keystone

A front-office decision platform for every MLB club: each decision traced through to wins, built on public data, with every model graded in public

Fifteen decision tools run from the winter to the dugout: a roster builder, a free-agent board, a contract studio, lineup and bullpen labs, and an ABS challenge coach. Underneath sits a Python and DuckDB pipeline over 87 public datasets, and an evidence wing that tests 67 hypotheses by written rules and publishes every verdict, the failures included. Where public data goes dark, Keystone says so rather than guessing.

The launch film: the winter plan, for any organization. Live at keystone-baseball.comFig. 02 of 06

03 · Care operations

Rebound

One board, one row per athlete. Providers plan, athletes follow, and the plan never falls out of sync between them

A triage strip decides who needs attention first; a pain-flagged skip outranks everything. Every session wears its track's color, so a week reads at a glance, and athletes check in from a PIN-guarded link: no app, no login. A practicing physical therapist put it through its paces, and her feedback shipped as set-scheme builders, supersets, and left and right sides.

The launch film: providers plan, athletes follow. Live at rebound.careFig. 03 of 06

04 · Consumer product

What Should We Make

The whole product is one promise, plan your family's meals in 60 seconds, and every screen supports it

Pick dinners, pair sides that quietly track your plate's color variety, and one tap produces an organized grocery list. Friendly, rounded, zero-jargon: designed for the least technical person in the household. In my own kitchen, an autonomous agent takes the list the rest of the way to a placed Instacart order.

The launch film: picked, paired, shopped. Live at whatshouldwemake.xyzFig. 04 of 06

05 · Ecosystem visualization

LUCI Explorer

Sam Spratt's LUCI universe as a heliocentric map: 929 tokens in orbit, every collector's belonging drawn as connection

Shipped as my vision for Sam's site overhaul after a phone call with the artist; he adopted the heliocentric model as a central part of his redesign. The 1/1 paintings sit at the center and each layer orbits outward in the order the world grew. The rings were engineered to render live on-chain data via Alchemy's NFT API, since retired to a snapshot to keep a community tool cheap to run.

LUCI Explorer. 929 tokens. One orbit. The live explorer: four rings of real token artwork around the ten LUCI paintings, with the collection overview panel
Hover any token and its collector's network lights up
The live explorer, as it renders today. Every dot is a real tokenFig. 05 of 06

06 · Market intelligence

Good Vibes Club Analytics

One honest, always-current picture of a collection's market: a dashboard that reads like a publication

A numbered editorial system (01 Price & Volume through 04 Collection Health) organizes the page, ETH/USD and time-range toggles persist, and every visualization exports a socially optimized graphic, branded and feed-sized, so any chart or sale is one tap from being shared.

Good Vibes Club Analytics. Live numbers, feed-ready. The live dashboard at goodvibesclub.info with the sale-card and Holding Period exports layered over it
A real export: branded, feed-sized, one tap
The live dashboard and its one-tap exports, exactly as the site renders themFig. 06 of 06

Process

How the work gets made

I

Design the foundation

Every build starts with the bones: the backend infrastructure, the data model, and a first-draft layout that respects both. It's the instinct I honed at Salesforce, where I owned the BI department's Tableau environment and standardized its templates, unifying our brand identity and eliminating project bloat. Define the design once, and every update arrives already matching it.

Receipt: one design system, six shipped products

II

Refine as you build

The first draft never survives the build; the design keeps tightening in code until it feels right. Information architecture, visual system, interaction, then production code with Claude Code and agentic workflows. Every case study above shipped without a handoff. Engineering isn't a blind spot for me: I'm happy to dive into the weeds, or build alongside other engineers.

Receipt: a practicing PT's feedback shipped as features

III

Let the data judge the work

Delivery is the start of the critique, not the end. At Cigna, A/B testing drove double-digit gains in customer satisfaction for complex-care populations; self-serve dashboards at Salesforce returned ~32 analyst-hours a week. The metric I design for: do people keep coming back? If not, it isn't finished yet.

Receipt: 67 Keystone hypotheses graded in public, failures included

30MLB front offices, one platform
67hypotheses graded in public
30,000+regulations, cited verbatim
27real care protocols, live
929tokens mapped in one orbit
14,346 ETHmarket history, tracked live

About

Analyst by training, builder by conviction

I'm Dan, based in Ohio and proud of it. I spent a decade leading analytics at Cigna, Salesforce, and Deloitte, studied Business Analytics at Harvard as the program's second-youngest graduate, and hold an MBA in Finance.

Now I build AI-natively, end to end. I direct agentic workflows in Claude Code to take a product from a blank folder to a deployed, data-backed release, across industrial compliance, pro sports, healthcare operations, consumer software, and Web3 communities. Every build is held to the bar I held dashboards to for a decade: does each number trace to a source, and do people come back? That's what I bring: the systems thinking of an analyst, the judgment of a designer, and the shipping habit of a builder.

Contact

Let's make something
people come back to

Email Dan LinkedIn Resume