PRODUCT DESIGN
AI INTEGRATION
FULL-STACK BUILD
PYTHON
OPENAI API
PRODUCT DESIGN
AI INTEGRATION
FULL-STACK BUILD
PYTHON
OPENAI API
PRODUCT DESIGN
AI INTEGRATION
FULL-STACK BUILD
PYTHON
OPENAI API
CASE STUDY — 2024 · SOLO BUILD
Ethic Flow: an
Mission-driven founders can't afford UX consultants. I built a fully automated pipeline that takes a website screenshot and a short survey, runs it through GPT-4, and emails a personalized audit report, in my own voice, within minutes. No human in the loop.
CASE STUDY — 2024 · SOLO BUILD
Ethic Flow: an
Mission-driven founders can't afford UX consultants. I built a fully automated pipeline that takes a website screenshot and a short survey, runs it through GPT-4, and emails a personalized audit report, in my own voice, within minutes. No human in the loop.
CASE STUDY — 2024 · SOLO BUILD
Ethic Flow: an
Mission-driven founders can't afford UX consultants. I built a fully automated pipeline that takes a website screenshot and a short survey, runs it through GPT-4, and emails a personalized audit report, in my own voice, within minutes. No human in the loop.
CONTEXT
Growth Ethic Experts
ROLE
Solo designer + engineer
BUILD TIME
1 week
STATUS
Shipped & live
CONTEXT
Growth Ethic Experts
ROLE
Solo designer + engineer
BUILD TIME
1 week
STATUS
Shipped & live
CONTEXT
Growth Ethic Experts
ROLE
Solo designer + engineer
BUILD TIME
1 week
STATUS
Shipped & live
◆ THE PIPELINE
I designed and built a fully automated audit system. A user fills out a Tally form, uploads screenshots of their website, and within minutes receives a detailed UX audit email written in my voice.
⊡
Survey + upload
Tally
→
⬡
Backend processing
Python · Render
→
◈
Audit generation
OpenAI API
→
✉
Delivery
Brevo
WHY TALLY
Native file upload + form in one
Tally handles image uploads alongside structured survey responses natively, no separate file storage layer needed. Non-technical founders could submit screenshots without friction.
WHY BREVO
Transactional email built for this
Brevo's transactional email API gave reliable programmatic delivery with formatting control. The audit needed to land in an inbox looking polished.
◆ THE PIPELINE
I designed and built a fully automated audit system. A user fills out a Tally form, uploads screenshots of their website, and within minutes receives a detailed UX audit email written in my voice.
⊡
Survey + upload
Tally
⬡
Backend processing
Python · Render
◈
Audit generation
OpenAI API
✉
Delivery
Brevo
WHY TALLY
Native file upload + form in one
Tally handles image uploads alongside structured survey responses natively, no separate file storage layer needed. Non-technical founders could submit screenshots without friction.
WHY BREVO
Transactional email built for this
Brevo's transactional email API gave reliable programmatic delivery with formatting control. The audit needed to land in an inbox looking polished.
◆ THE PIPELINE
I designed and built a fully automated audit system. A user fills out a Tally form, uploads screenshots of their website, and within minutes receives a detailed UX audit email written in my voice.
⊡
Survey + upload
Tally
⬡
Backend processing
Python · Render
◈
Audit generation
OpenAI API
✉
Delivery
Brevo
WHY TALLY
Native file upload + form in one
Tally handles image uploads alongside structured survey responses natively, no separate file storage layer needed. Non-technical founders could submit screenshots without friction.
WHY BREVO
Transactional email built for this
Brevo's transactional email API gave reliable programmatic delivery with formatting control. The audit needed to land in an inbox looking polished.
◆ THE PROBLEM I WAS SOLVING
Mission-driven founders, nonprofits, wellness businesses, ethical tech startups often have real UX problems dragging down their conversion, but can't afford a $5k consulting sprint to find out what they are. I wanted to build something that democratized that expertise: instant, actionable, and in a voice that felt like a real advisor, not a generic AI checklist.
The AI wasn't just generating text, I engineered the prompt so the output would read in my specific analytical voice, structured the way I would actually write a consulting deliverable. That meant the output had to feel personal even though the pipeline was fully automated.
◆ THE PROBLEM I WAS SOLVING
Mission-driven founders, nonprofits, wellness businesses, ethical tech startups often have real UX problems dragging down their conversion, but can't afford a $5k consulting sprint to find out what they are. I wanted to build something that democratized that expertise: instant, actionable, and in a voice that felt like a real advisor, not a generic AI checklist.
The AI wasn't just generating text, I engineered the prompt so the output would read in my specific analytical voice, structured the way I would actually write a consulting deliverable. That meant the output had to feel personal even though the pipeline was fully automated.
◆ THE PROBLEM I WAS SOLVING
Mission-driven founders, nonprofits, wellness businesses, ethical tech startups often have real UX problems dragging down their conversion, but can't afford a $5k consulting sprint to find out what they are. I wanted to build something that democratized that expertise: instant, actionable, and in a voice that felt like a real advisor, not a generic AI checklist.
The AI wasn't just generating text, I engineered the prompt so the output would read in my specific analytical voice, structured the way I would actually write a consulting deliverable. That meant the output had to feel personal even though the pipeline was fully automated.
◆ HOW IT CAME TOGETHER
Day 1
Problem definition + tool selection
Mapped the user journey from "founder with a leaky website" to "has an audit in their inbox." Selected Tally for its file upload capability, Render for serverless Python, Brevo for delivery.
Days 4–5
Prompt engineering + voice calibration
Iterated on the system prompt until the GPT-4 output consistently read in my analytical voice; specific, structured, and actionable rather than generic. This was the hardest part.
Days 2–3
Backend pipeline build
Built the Python backend on Render webhook receiver from Tally, image parsing, OpenAI API integration with a structured system prompt engineered to produce audit-style output.
Days 6–7
Email formatting + end-to-end testing
Wired Brevo, formatted the email template so audit output rendered cleanly across clients, tested the full pipeline end-to-end. Shipped the landing page. Live.
◆ HOW IT CAME TOGETHER
Day 1
Problem definition + tool selection
Mapped the user journey from "founder with a leaky website" to "has an audit in their inbox." Selected Tally for its file upload capability, Render for serverless Python, Brevo for delivery.
Days 2–3
Backend pipeline build
Built the Python backend on Render webhook receiver from Tally, image parsing, OpenAI API integration with a structured system prompt engineered to produce audit-style output.
Days 4–5
Prompt engineering + voice calibration
Iterated on the system prompt until the GPT-4 output consistently read in my analytical voice; specific, structured, and actionable rather than generic. This was the hardest part.
Days 6–7
Email formatting + end-to-end testing
Wired Brevo, formatted the email template so audit output rendered cleanly across clients, tested the full pipeline end-to-end. Shipped the landing page. Live.
◆ HOW IT CAME TOGETHER
Day 1
Problem definition + tool selection
Mapped the user journey from "founder with a leaky website" to "has an audit in their inbox." Selected Tally for its file upload capability, Render for serverless Python, Brevo for delivery.
Days 2–3
Backend pipeline build
Built the Python backend on Render webhook receiver from Tally, image parsing, OpenAI API integration with a structured system prompt engineered to produce audit-style output.
Days 4–5
Prompt engineering + voice calibration
Iterated on the system prompt until the GPT-4 output consistently read in my analytical voice; specific, structured, and actionable rather than generic. This was the hardest part.
Days 6–7
Email formatting + end-to-end testing
Wired Brevo, formatted the email template so audit output rendered cleanly across clients, tested the full pipeline end-to-end. Shipped the landing page. Live.
◆ FULL STACK
◆
Tally
Survey + image intake
⬡
Python
Backend logic
⊡
Render
Hosting + webhooks
◈
OpenAI API
GPT-4 audit generation
Brevo
Transactional email
◆ FULL STACK
◆
Tally
Survey + image intake
⬡
Python
Backend logic
⊡
Render
Hosting + webhooks
◈
OpenAI API
GPT-4 audit generation
Brevo
Transactional email
◆ FULL STACK
◆
Tally
Survey + image intake
⬡
Python
Backend logic
⊡
Render
Hosting + webhooks
◈
OpenAI API
GPT-4 audit generation
Brevo
Transactional email
◆ VIEW THE SOURCE
The full pipeline is open source. The code covers the webhook receiver, OpenAI API integration, PDF generation, and Brevo email delivery.
⬡ ethic-flow
PUBLIC
Fully automated UX audit pipeline: Tally intake → Python/Render backend → GPT-4o audit generation → Brevo email delivery. Shipped solo in one week.
◆ VIEW THE SOURCE
The full pipeline is open source. The code covers the webhook receiver, OpenAI API integration, PDF generation, and Brevo email delivery.
Fully automated UX audit pipeline: Tally intake → Python/Render backend → GPT-4o audit generation → Brevo email delivery. Shipped solo in one week.
◆ VIEW THE SOURCE
The full pipeline is open source. The code covers the webhook receiver, OpenAI API integration, PDF generation, and Brevo email delivery.
Fully automated UX audit pipeline: Tally intake → Python/Render backend → GPT-4o audit generation → Brevo email delivery. Shipped solo in one week.
"I built this in a week during a hard personal moment, which is also proof of something. When I care about a problem, I ship. The tool worked. The pipeline ran. And it showed me I can build things, not just research them."
"I built this in a week during a hard personal moment, which is also proof of something. When I care about a problem, I ship. The tool worked. The pipeline ran. And it showed me I can build things, not just research them."
"I built this in a week during a hard personal moment, which is also proof of something. When I care about a problem, I ship. The tool worked. The pipeline ran. And it showed me I can build things, not just research them."