NutriLens is an AI-driven mobile application concept that eliminates manual food logging by enabling users to photograph meals, receive instant macro estimates, and get intelligent meal recommendations to achieve daily nutrition goals.
The project addresses a real-world problem: existing calorie trackers (MyFitnessPal, etc.) suffer from tedious manual logging, poor accuracy for local cuisines, and lack of proactive guidance.
| Artifact | Description |
|---|---|
| Business Statement | Organization background & client problem/opportunity |
| Feasibility Analysis | Technical, Operational, Economic, Schedule (rated 1-10) |
| SMART Goals | 5 measurable success criteria (accuracy, latency, onboarding, retention, rating) |
| Project Deliverables | 15 deliverables with acceptance criteria |
| Project Scope | Clear inclusions, exclusions, and 8 phases |
| Agile User Stories | 22 stories (New User, Registered User, Admin) |
| Assumptions, Constraints, Dependencies | 7 assumptions, 5 constraint categories, 7 dependencies |
| Stakeholder Analysis | Roles, responsibilities, signing authorities |
| Expert Consultation | Mr. Mohammad Kashif (ML/Computer Vision expert) |
| Work Breakdown Structure | 215+ activities with 3-level numbering |
| MS Project Plan | Gantt Chart, PERT Network, critical path (~239 days) |
| Risk Management | 5 critical risks with mitigation plans |
| Time-Phased Budget | Resource Sheet, Cost Sheet, Summary (USD/PKR) |
| Role | Key Responsibility |
|---|---|
| Project Manager | Planning, budget, risk, stakeholder communication |
| ML Engineer | MobileNetV2 model training, TFLite, Gemini integration |
| Backend Developer | Flask API, Firebase, A* meal planner |
| Flutter Developer | Mobile frontend, camera, UI screens |
| QA Engineer | Testing, UAT, bug tracking |
| UI/UX Designer | Wireframes, mockups, style guide |
| DevOps Engineer | CI/CD, deployment, monitoring |
| Phase | Duration | Key Activities |
|---|---|---|
| 1. Initiation & Planning | Mar 2 – Apr 23 (38d) | Project charter, requirements, WBS, budget |
| 2. Design | Mar 2 – May 11 (49d) | UI/UX, database, API design |
| 3. ML Development | Apr 24 – Jun 17 (35d) | Model training, TFLite conversion |
| 4. Backend Development | May 12 – Aug 21 (68d) | Flask API, Gemini, A* planner |
| 5. Frontend Development | Jun 18 – Jul 28 (27d) | Flutter screens, onboarding, camera |
| 6. Testing & QA | Aug 24 – Oct 2 (29d) | Unit, integration, UAT |
| 7. Deployment | Jul 29 – Aug 18 (14d) | Production setup, Play Store |
| 8. Project Closure | Oct 5 – Oct 29 (19d) | Review, handover, documentation |
Total Duration: 239 working days (~8 months)
Critical Path: Planning → Design → ML → Backend → Frontend → Testing → Deployment → Closure
| Metric | Amount |
|---|---|
| Total Labor Cost | (Calculated in Excel) |
| Contingency (10%) | (Calculated) |
| Overhead (15%) | (Calculated – covers hosting, APIs, tools) |
| Grand Total | (Calculated) |
Exchange Rate: 1 USD = 278 PKR (March 2026 assumption)
| Risk | Severity | Mitigation |
|---|---|---|
| Low ML Model Accuracy | Critical | Top-3 predictions, manual override |
| Gemini API Downtime/Cost | High | Fallback, billing alerts |
| Backend Cold Starts | High | Keep-alive ping, upgrade tier |
| Key Resource Unavailability | High | Documentation, cross-training |
| Play Store Rejection | High | Early registration, policy compliance |
Name: Mohammad Kashif
Domain: Machine Learning, Deep Learning, Computer Vision
Role: External Technical Advisor: reviewed WBS, feasibility, and risk strategies.
- ≥75% Top-1 classification accuracy
- <2 seconds end-to-end latency
- ≥85% onboarding completion
- ≥40% first-week retention
- ≥4.0/5.0 meal planner rating