Data Analyst · AI Builder · Founder

I like building useful things from messy real-world problems — especially where data, AI, product thinking, and human behavior overlap.

My background is in analytics, statistics, and data visualization. Lately, I’ve been taking that same way of thinking into product builds: turning handwritten appointment books into business intelligence, exploring face-to-face translation on new hardware, building workforce matching tools, and creating reusable security standards for AI-assisted development.

I care about products that are practical, explainable, and actually useful to the person on the other side of the screen.

A menstrual cycle you can see, hear, and feel.

Mensis is an interactive data artwork that turns cycle research into a tactile experience. Visitors move through a menstrual cycle using a ring of textured moon phases while a glass fills with an illustrative flow model, phase-aware sound translates the data into music, and an optional AI explanation connects a selected cycle day and symptoms to the displayed phase in plain, non-diagnostic language.

The experience distinguishes research findings, clinical guidance, and modeled data directly in the interface. It uses mcPHASES and ACOG sources while treating the moon as a cultural metaphor—not a biological clock.

Built with: Cala, Krea, Lovable, React/TypeScript, Web Audio, Mistral AI
Privacy: No personal cycle data is stored. Educational only; not medical advice.
Live experience: https://pixel-moon-canvas.lovable.app

My interactive toolkit for the tools I use to build, analyze, automate, and ship.

Jada Data Stack is a centralized resource hub for the apps, platforms, and developer tools I use across data analysis, AI building, automation, product development, and deployment. It organizes the stack by what each tool is best for so I can quickly choose the right tool for the job and share the same resource with other builders.

Built with: Lovable, TanStack Start, TypeScript
Live site: https://jada-data-stack.lovable.app

Appointment-book intelligence for independent hairstylists.

COMBED reads photos of a handwritten salon appointment book, structures the appointments, calculates business metrics, adds local-market context, and turns the results into a plain-language 20-day action plan.

Built from a problem close to home: helping my mom understand how to keep her chair full while spending her time more strategically.

Built with: JavaScript, Node/Express, vision + LLM reasoning, Apify
Recognition: Built at Claude Build Day and received a special project award.

Tell me in your language.

Dime is a face-to-face translation experience designed around the iPhone Duo form factor. Instead of passing one phone back and forth, each person gets their own side of the conversation.

The hackathon MVP focuses on English ↔ Español communication in a healthcare setting, with accessibility, clarification, recap, and interpreter escalation built into the product idea.

Built with: Swift, SwiftUI, Swift Package Manager, RevenueCat test tooling

A skills-first bridge between union dispatches.

BenchBridge explores workforce matching by starting with a worker’s real skills, certifications, experience, location, and requirement gaps instead of just a job title.

The prototype connects workers with programs, employers, project signals, and realistic next steps based on the skills and requirements that matter for each opportunity.

Built with: TypeScript/JavaScript, structured public-source research
Live demo: https://benchbridge-liard.vercel.app

My security baseline for AI-assisted app building.

I move fast when I build, but I do not want “move fast” to become an excuse for careless security decisions.

JadaSecure is the reusable security standard I bring into AI-assisted projects so auth, access control, secrets, privacy, and launch checks are part of the build from the beginning.

A refreshed statistical case study using anonymized youth gymnastics competition data. The project focuses on effect sizes, uncertainty, exact permutation testing, robust standard errors, responsible missing-data choices, and knowing when the evidence does not support a dramatic conclusion.

Live analysis: https://jadamylan.github.io/balancing-beam-of-performance/

An interactive rebuild of a 2022 graduate statistics project exploring the chemical properties associated with Portuguese Vinho Verde quality — plus a 2026 look at how I would approach the same question with a more mature analytical stack.

Live analysis: https://jadamylan.github.io/wine-quality-analysis/

I opened Mistral Vibe PR #1028 — Add /usage for cross-session cost and token reporting.

The contribution adds a read-only usage aggregator over local session logs plus a /usage workflow for reviewing cost and token usage across 7 days, 30 days, or all history, with optional CSV output and a one-line model-generated insight.

The PR is currently open. The implementation includes dedicated tests and documentation updates.

I enjoy contributing to tools I actually use because it forces me to move from:

“I wish this existed.”

to:

“Cool. Let me see if I can build it.”

Jasiri Coder — a math and coding tutoring business I co-founded. I’m building the product layer around it, including student skill diagnostics, tutor workflows, and tools that help us scale beyond one-on-one tutoring without losing the human handoff.

A few patterns show up in almost everything I build:

• Start with the person, not the technology.
• Make the numbers explainable.
• Prototype quickly, but keep the guardrails real.
• Use AI as a builder and reasoning partner — not as an excuse to skip judgment.
• Build from problems I have actually seen people deal with.

I’m especially interested in analytics, AI products, workforce tools, education, small-business technology, and creative hackathon builds.

• Jada Data Stack — interactive toolkit for my data, AI, automation, and build stack
• Mensis — interactive menstrual-cycle data artwork
• BenchBridge — skills-first workforce matching
• Balancing Beam of Performance — statistical case study
• What Makes a Good Wine? — interactive statistics portfolio

San Francisco Bay Area · GitHub