Why I Built Eatomate
I'm Rohith, a Senior Software Engineer who has worked at Meta, HubSpot, and HackerRank. I got frustrated with nutrition tracking apps that don't work, so I left my FAANG job to build the solution I wanted—full-time.
The Problem I Faced
Like many engineers optimizing their health, I tried tracking nutrition with existing apps. 2 minutes per meal of manual logging. Food databases with wildly inaccurate entries. Generic "sambar" that doesn't match how my grandmother makes it.
I lasted 3 weeks before giving up. The accuracy was terrible, and the time investment wasn't sustainable.
I realized this is a solvable engineering problem. A cheap kitchen scale can measure portions in grams with statistical confidence. Database matching across barcodes, recipes, and aliases can identify what you're eating, and physics (mass conservation) can converge to ~95% weekly accuracy through reconciliation. The system learns your patterns over time.
So I built Eatomate.
The Breakthrough
I realized that mass conservation (consumed = purchased - remaining) combined with the Central Limit Theorem means you don't need to weigh every meal.
You only need to:
- Weigh 3 portions to learn your Gaussian distribution (mean ± std dev)
- Single-tap meals from predictions 90% of the time (1 second)
- Let expiry dates auto-reconcile 70-80% of pantry items
- Reweigh ~5 slow-expiring staples per week (60 seconds)
Result: 1.6 minutes per week, 95% accuracy
vs. competitors: 42-63 minutes/week, 50-70% accuracy
This is why Eatomate is a physics breakthrough, not just another tracking app.
Built by Someone Who Gets It
Built Scalable Systems
At Meta, I worked on Workplace's billing and monetization system, handling subscription lifecycle, prorations, and invoice generation for enterprise customers. At Media.net, I developed a highly modular Java crawling library for web crawling infrastructure.
→ I've built production systems that handle large-scale data processing and user interactions.
Full Stack Expertise
I've worked across the full stack with backend (Java, Python, Hack) and frontend (React, BackboneJS) technologies. At Meta, I worked as a full stack engineer (65% backend, 35% frontend). At Instamojo, I started as a backend developer and moved to full stack development.
→ I can build the entire product myself, ensuring quality end-to-end.
Data Systems Experience
At HubSpot, I worked on real-time data syncing applications building a CRM sync validator. At Meta, I worked on billing systems with complex data reconciliation requirements.
→ Nutrition tracking requires reconciling messy real-world data (receipts, portions) into accurate insights.
Startup Experience
At Instamojo, I worked on features touching all parts of the product and handled a portion of customer support for technical issues. At HackerRank, I worked on their enterprise product building critical modules.
→ I can move quickly and iterate based on user feedback.
Background
The Vision
Nutrition tracking should take 1.5 seconds per meal, not 2-3 minutes.
It should be ~95% accurate after 4 weeks through calibration + reconciliation, learning your recipes over time — not relying on generic database guesses.
It should take 1.6 minutes per week total, not 42-63 minutes.
That's what I'm building with Eatomate.
Transparency: I'm not a nutritionist or dietitian. Eatomate provides accurate tracking data—interpretation should come from qualified professionals. I'm building the measurement tool, not prescribing diets.
Start Your Journey Today
Try research-grade nutrition tracking with a free 14-day trial.