Founder Log Week #1 – From Skeleton UI to Real App

- 4 minutes read - 779 words

🧠 Weekly Overview

This was the first true build week for Menuvivo. I started with nothing but a vision and a goal: prove this idea can be real. In just nine days, I designed and deployed the foundational UI, implemented simulated flows (from fridge photo to recipe generation), and built a fully automated deployment pipeline for frontend. I was also advanced with a backend deployment pipeline, but I didn’t manage to finalize it by the end of the week. I also experienced some of the emotional rollercoaster every founder knows: excitement, doubt, frustration, joy.

I also decided to document my founder journey from the first day of Menuvivo implementation, and this is the first post. I’ll publish a new weekly summary every Sunday afternoon. Stay tuned.

🗓 Daily Execution Summary

Mar 22 (Sat)

Set up React + MUI app, built the initial ingredient list component with localStorage support. Polished basic theme with Menuvivo colors. Deployed to AWS S3 + CloudFront manually. CI/CD for staging started taking shape.

Mar 23 (Sun)

Fixed S3 upload bug and finished GitLab pipeline: each MR gets its own staging env, and main branch deploys to production. Application is now live with working demo flow. Felt unsure whether infra investment was worth the time, but proud of clean setup.

Mar 24 (Mon)

Implemented photo-taking flow in the UI using a file input that works across mobile and desktop. Simulated the result by showing a hardcoded ingredient list after upload. Deployed to staging. Balanced work with family duties.

Mar 25 (Tue)

Completed the recipe generation mock flow. The UI now simulates taking a photo, extracting ingredients, and suggesting meals. Clean mobile-first layout and real momentum felt. Still managed it despite a doctor appointment disrupting the day.

Mar 26 (Wed)

Mocked AI image recognition: upload photo, delay, then inject 3–5 fake ingredients with visual tags. Highlighted which ingredients are used in the suggested recipes. Loved the result. First glimpse of what Menuvivo could feel like.

Mar 27 (Thu)

Bug fixing day. Editing ingredient items was broken. AI struggled to help. Switched models mid-debug. Eventually solved it but lost time. Realized that AI helps most with new code, not bug fixing. Didn’t get to polish layout or build feedback prompt.

Mar 28 (Fri)

Huge win: fully working bottom nav bar across mobile, tablet, and desktop. App feels like a real product now. Feedback chat bubble UI added (no backend yet). All primary screens are stubbed and deployed. Felt euphoric.

Mar 29 (Sat)

Designed the backend infra for collecting feedback and wrote a reusable ADR with Archie (my AI architect). Spent time setting up the new service and its build config. Didn’t ship direct business value today, but delivered high-leverage architecture.

Mar 30 (Sun)

Focused on CI/CD for the backend. Built and tested GitLab pipeline for Lambda deployment. Still facing build issues but, generalized pipeline infra so I can reuse it later. Progress was slower due to poor sleep and the time shift.

🏆 Wins of the Week

  • React + MUI UI with inventory, photo input, and recipe suggestion flows
  • Ingredient recognition simulation logic with image cropping preview
  • Bottom nav bar for mobile/tablet/desktop
  • Feedback UI with chat bubble and modal (frontend only)
  • All UI deployed to S3 + CloudFront with GitLab CI/CD, see app.menuvivo.com
  • GitLab CI/CD pipeline for staging/production frontend
  • Generalized GitLab pipeline infra for backend (AWS Lambda)
  • Architecture Decision Record for reusable backend stack

✍️ Lessons & Reflections

  • Building “AI Agents” (prompts designed to solve specific problems) is a game changer, this week I created the following agents:
    • Archie (AI architect) helps with architecture decisions and ADRs
    • Tim (AI business analyst) helps with product flows and user stories
  • Aider powered by Sonnet 3.7 or Deepseek R1 is quite effective in implementing features that are well-defined by Tim
  • Debugging and fine-tuning are still mostly manual
  • Having a rock-solid deployment loop from day one saves time and stress later
  • Frontend polish matters more than I expected: when the app looked real, I believed in it more
  • Sleep and diet directly affect execution – a massive impact on energy and clarity. Trying to eat better and have a solid sleep routine: wake up at 5:00 and go to bed by 21:30.

🎯 Plan for the next week

Next week is the second half of my MVP sprint. I plan to:

  • Finalize feedback backend and test full flow (UI → Lambda → S3)
  • Start user validation interviews and sign-ups
  • Connect the feedback feature form app to waitlist newsletter
  • Write a blog post about how I started Menuvivo
  • Start recipe search and filtering feature

This week I proved I can build and ship fast. Now I must validate, iterate, and prepare for launch.