From Hard-Coded to Smart: Building a Personalized Nutrition Calculator

It started with a simple question in my community chat.

Eric, a user who had been successfully using the app for a month, reached out: "I really like it (have been losing weight, yay!), but I was wondering—is there a way to tell the AI to adjust the protein/carb targets? They're set a bit higher than I'd like right now."

That message was the spark. It highlighted a limitation in my early MVP: I was using a "one size fits all" approach. I had hard-coded defaults that worked for "average" cases but failed to account for the diverse needs of my actual users.

I realized it was time to graduate from hard-coded constants to a smart, dynamic nutrition engine. Here is the story of how I built the "Reset to Recommended Values" feature, the math behind it, and the challenges I solved along the way.

The Old Way: Hard-Coded & Rigid

In the beginning, my focus was on getting the core tracking mechanics right. For targets, I took a shortcut: standard ratios (30% Protein, 40% Carbs, 30% Fat) applied to a generic calorie baseline.

If you were an athlete, a sedentary office worker, or someone trying to gain muscle, you basically got the same starting numbers. You could manually edit them, but the app didn't help you find the right numbers.

Building the "Smart" Calculator

To make the app truly helpful, I needed to calculate targets based on the individual. I turned to the Mifflin-St Jeor equation.

In the world of nutrition science, Mifflin-St Jeor is widely considered the "gold standard" for estimating Basal Metabolic Rate (BMR). In simple terms, BMR is the energy your body burns just to keep the lights on—breathing, circulating blood, and cell repair—while completely at rest.

But nobody stays completely at rest. So, I take that BMR and multiply it by an "Activity Factor" to get your Total Daily Energy Expenditure (TDEE):

• Sedentary: BMR × 1.2
• Lightly Active: BMR × 1.375
• Moderately Active: BMR × 1.55
• Very Active: BMR × 1.725
• Extra Active: BMR × 1.9

This gave me a much better baseline: the amount of calories you need to maintain your current weight given your lifestyle.

Understanding User Goals

Knowing how many calories you burn is only half the battle. Most of my users have a specific goal in mind.

I implemented a "smart" goal parser that listens to what the user wants. Instead of forcing users to select from a rigid dropdown, I analyze their intent:

• "Lose weight": I subtract 500 kcal from their TDEE. This creates a moderate deficit designed for sustainable weight loss (roughly 1 lb per week).
• "Slow loss": I apply a smaller 250 kcal deficit.
• "Gain muscle": I add a 300 kcal surplus to fuel growth.

I then distribute these calories into macros that make sense: adequate protein for muscle retention, balanced fats for hormonal health, and carbs for energy. I even added a recommendation for fiber (14g per 1000 calories) to ensure diet quality, not just quantity.

Solving the "Safety Floor" Problem

Engineering often reveals edge cases that theory misses. I encountered a critical issue during testing with sedentary users.

I have a "safety floor" in my code to prevent the app from recommending dangerously low calorie intakes. Initially, this floor was set to BMR * 1.2.

The problem? A sedentary user's TDEE is also BMR * 1.2.

When a sedentary user wanted to lose weight, my math would try to subtract 500 calories. But the safety floor would immediately kick in and push the target back up to BMR * 1.2. The result: the app refused to give them a calorie deficit!

I fixed this by lowering the absolute safety floor to BMR (the raw survival calories). This created a safe "buffer zone" where sedentary users could still have a healthy deficit without dipping below their body's basic needs.

Conclusion

This journey wasn't just about adding a calculator; it was about making the app listen and adapt.

By combining the precision of the Mifflin-St Jeor equation with goal-oriented logic, I've turned a static tracker into a dynamic guide. Now, when a user asks, "Can I adjust my targets?", the answer isn't just "Yes, you can edit them manually"—it's "Yes, let's calculate exactly what you need."

The new Smart Targets feature is available in the app now. Go to your Profile Settings to try the "Reset to Recommended Values" button.