Proposal

Menu systems,
AI-driven.

A structured system for scalable, automated menu generation — combining nutrition science with AI-powered operations.

Prepared for
Mikhail Khardikov
Prepared by
Antonio Yglesias & Laura Murillo
Date
April 2026
Scroll
About
Who we are

We're Antonio Yglesias and Laura Murillo — a team who built and ran La Oliva Verde, a healthy meal prep operation with 10 locations and a centralized production kitchen. We managed 200+ standardized recipes in an ERP with in-house macro calculations, cost tracking, and menu standardization across all locations.

Laura Murillo is a Food Technologist with a Master's in Nutrition & Metabolism, HACCP certified, 15+ years in formulation and recipe development. Antonio Yglesias managed business operations and now builds AI-driven automation systems — structured data pipelines, rule-based generation engines, and workflow orchestration.

Understanding
What you need

A team to own three things: structuring your 700-recipe database (currently split between Nutribot and Excel), delivering weekly menus across all plans and goals, and developing new recipes to maintain variety.

You also want to move toward AI-assisted menu generation. That's where we're different from a traditional hire.

Cost volatility is a real challenge. Ingredient prices shift and menus need to adapt without breaking nutritional or variety standards. Our system handles this — when costs change, we identify affected recipes, flag menus that exceed thresholds, and suggest substitutions that maintain nutrition and variety while optimizing for the new cost reality.

Engagement
Three phases
Phase 1 — Database Setup
Weeks 1–4
$2,400
fixed fee
  • Access current data from Nutribot and Excel — audit completeness and quality
  • Design tagging schema in Airtable with structured fields, dropdowns, and form views
  • Migrate first 50 recipes as proof of concept — checkpoint with your team
  • Migrate and tag all remaining recipes — validate nutritional data and fill gaps
  • Set up automated weekly CSV export to Google Drive
  • Deliverable: complete, validated, AI-ready recipe database
Payment: $1,200 at project start — $1,200 on delivery and approval
Phase 2 — Menu Delivery
Weeks 5–8
$1,650
fixed fee
  • Weekly menu sets for each of the 9 current plans — one complete deliverable per week covering all plans
  • Database maintenance — cost updates, availability changes, new tags
  • AI system development and optimization
  • Documentation of all menu composition rules
$825 at phase start — $825 on delivery and approval
Phase 3 — Steady State
Month 3+
$1,450
per month
  • Weekly menu sets for each of the 9 current plans
  • Up to 18 new recipes per month — fully costed, macro-calculated, tagged, and prioritized to add variety where plans need it most
  • Database maintenance and ongoing updates
  • Cost monitoring and optimization — flag affected recipes, suggest substitutions, adjust plans to stay within budget
  • AI system optimization and continuous improvement
  • One seasonal or promotional menu per quarter
50% at the start of each month — 50% at month end upon delivery
Not included — priced separately
  • Adding new meal plans or goal categories beyond the current set
  • Major recipe database reformulations (20%+ of existing recipes)
  • Custom integrations with Nutribot or other systems
  • Procurement list generation or supplier management
System
How it works
Layer 1

Recipe Database

Airtable as the single source of truth. Recipes, costs, and tags can be updated directly with structured fields that prevent data entry errors. Full version history and audit trail built in.

Layer 2

Automated Export

Weekly CSV snapshot to Google Drive. No manual step, no version conflicts. Each snapshot doubles as a backup of the database state.

Layer 3

Rules Engine

Menu composition rules codified as adjustable parameters. Protein rotation, cost caps, variety, macro tolerances. Change a parameter, not 27 menus.

Layer 4

AI Generation

Pre-filters eligible recipes deterministically, then AI composes menus satisfying all constraints simultaneously. Every output validated before delivery.

Timeline
What to expect
Week 2
Structured database with 50 sample recipes migrated. Checkpoint to validate schema.
Week 4
Complete database delivered. Phase 1 done.
Week 6
First weekly menus delivered. Composition rules fully documented.
Week 8
AI-assisted menu generation running with full validation.
Month 3+
Steady state. AI-generated menus, recipe development active, cost optimization running.
Team
Roles

Laura Murillo

Nutrition & Food Science
  • Recipe development
  • Nutritional validation
  • Macro calculations
  • Menu quality review
  • Data quality ownership

Antonio Yglesias

Operations, Data & AI
  • Database architecture
  • AI system development
  • Rules engine
  • Automation pipeline
  • System maintenance
Terms
How we work together

Ownership & Handover

Everything we build is yours — the database, the rules, the prompts, the configurations, the documentation. If the engagement ends for any reason, we provide a complete handover package so your team can operate the system independently.

Scope Review

Quarterly review to assess whether deliverables, scope, or pricing need adjustment. If you add plans or goals, we propose an updated fee before starting. Either side can propose changes — no surprises.

Assumptions

This proposal is based on the following working assumptions. If any differ significantly from reality, we'll flag it during Phase 1 and discuss adjustments before proceeding.

  • Existing recipe data can be exported or shared in a tabular format (Excel, CSV, or similar) from both your Excel files and Nutribot
  • The majority of recipes already include nutritional values and cost data. We expect to fill gaps and standardize formatting, but our Phase 1 scope does not include calculating nutritional profiles from scratch for a significant portion of the database. If more than approximately 15% of recipes require full nutritional analysis, we'll flag this early and propose an adjusted scope
  • Menu composition rules exist today, even if informal or undocumented. We'll work with your team to capture and codify them
  • Your team has access to Google Workspace or can access Google Drive for file sharing
  • One point of contact on your side is available for weekly check-ins during Phase 1 and periodic feedback during steady state
Antonio Yglesias & Laura Murillo
Nutrition / Operations / AI-Driven Menu Systems
April 2026
Confidential