Igniters-Week2: digital twin

community_contributions/mugaoprodweek1/GeneInsight.md

@@ -0,0 +1,113 @@
+## 🧬 GeneInsight Studio
+
+I built a biotech-focused AI application that transforms raw genomic and experimental notes into structured biological insights, hypotheses, and stakeholder-ready summaries.
+
+The goal is to reduce the cognitive load in interpreting complex bioinformatics outputs while improving clarity, speed, and communication across research and clinical teams.
+
+**Live App**
+👉 https://gene-insight-studio-4xh4.vercel.app/
+
+<br />
+
+![Next.js](https://img.shields.io/badge/Next.js-16-black?logo=nextdotjs)
+![FastAPI](https://img.shields.io/badge/FastAPI-Backend-009688?logo=fastapi)
+![TypeScript](https://img.shields.io/badge/TypeScript-Frontend-3178C6?logo=typescript)
+![Python](https://img.shields.io/badge/Python-API-3776AB?logo=python)
+![OpenAI](https://img.shields.io/badge/AI-OpenAI-412991?logo=openai)
+![Docker](https://img.shields.io/badge/Deployment-Docker-2496ED?logo=docker)
+
+---
+
+## 🧬 Core Features
+
+* AI-powered interpretation of genomic and experimental data using OpenAI.
+* Structured output with three sections:
+
+  * **Research Summary** — clear biological interpretation of findings.
+  * **Hypotheses & Next Experiments** — actionable scientific next steps.
+  * **Regulatory / Stakeholder Summary** — simplified explanation for non-technical audiences.
+* Real-time streaming responses (Server-Sent Events) for immediate feedback.
+* Experiment input system:
+
+  * Sample type (RNA, DNA, tissue, etc.)
+  * Free-form experimental notes
+* Modular frontend architecture for scalability and clean UX.
+* Backend prompt engineering tailored for genomics and molecular biology use cases.
+
+---
+
+## Example Use Cases
+
+* RNA-seq differential gene expression analysis
+* Variant interpretation from DNA sequencing
+* Infectious disease genomics (antimicrobial resistance)
+* CRISPR experiment analysis
+* Pharmacogenomics and drug response insights
+
+---
+
+## Motivation
+
+Biological data is often:
+
+* fragmented
+* difficult to interpret
+* hard to communicate across teams
+
+GeneInsight Studio explores how AI can bridge the gap between:
+
+**Raw experimental data → Structured insight → Actionable decisions**
+
+This is particularly relevant for:
+
+* Biotech startups in Europe working on precision medicine
+* Research environments handling large-scale omics data
+* Resource-constrained healthcare systems where expertise is limited
+
+---
+
+## 🏗 Architecture Overview
+
+**Frontend (Next.js / React / TypeScript)**
+
+* User interface
+* Experiment input
+* Streaming output rendering
+
+**Backend (FastAPI / Python)**
+
+* Prompt construction
+* OpenAI API integration
+* SSE streaming responses
+
+**AI Layer**
+
+* Domain-specific prompt engineering for genomics reasoning
+
+---
+
+## ⚙️ Key Technical Highlights
+
+* **Server-Sent Events (SSE)** for real-time AI output streaming
+* **Domain-aware prompt design** for biological interpretation
+* **Separation of concerns** between UI and AI processing layers
+* **Environment-based configuration** using `.env` and `python-dotenv`
+* **Container-ready architecture** using Docker
+
+---
+
+## Deployment
+
+* Frontend deployed on **Vercel**
+* Backend containerized with **Docker** and deployable to AWS App Runner
+
+---
+
+## Repository
+
+The complete application is available at:
+https://github.com/Mugao-joy/geneInsight-Studio
+
+For full setup instructions, architecture details, and local development steps, see the main `README.md` in the repository.
+
+---

community_contributions/mugaoweek2prod.md

@@ -0,0 +1,220 @@
+# Digital Twin — AI Personal Assistant
+
+An AI-powered digital twin that represents me online. Visitors can interact with a conversational assistant that responds as me, using contextual knowledge and conversational memory. The system is built with FastAPI and deployed on AWS.
+
+---
+
+## 🔗 Project Repository
+https://github.com/Mugao-joy/digital-twin
+
+---
+
+## 🌐 Deployment URL
+
+Frontend (S3 Static Website):
+http://twin-frontend-yourname.s3-website-us-east-1.amazonaws.com/
+backend : https://ymw4zypayc.us-east-1.awsapprunner.com/
+
+> ⚠️ Note: This is an S3 static website endpoint. In a production setup, this would ideally be served via CloudFront for HTTPS and improved performance.
+
+---
+
+## Overview
+
+The project consists of three main parts:
+
+### Backend (FastAPI + AI)
+- Handles chat sessions and memory
+- Builds a system prompt to simulate a digital twin
+- Integrates with an AI model for responses
+- Uses **Mangum** to run FastAPI on AWS Lambda (initial approach)
+
+### Frontend
+- Simple chat interface
+- Sends user messages to backend API
+- Displays responses in real-time
+- Hosted on AWS S3
+
+### Infrastructure
+- **AWS Lambda** → backend execution (initial approach)
+- **API Gateway** → HTTP routing
+- **S3** → frontend hosting
+- **App Runner** → final backend deployment
+
+---
+
+## System Flow
+User → Frontend (S3) → API Gateway / App Runner → Backend → AI → Response
+
+---
+
+## Key Features
+
+- Conversational AI assistant
+- Session-based memory
+- Serverless / containerized backend
+- Lightweight frontend
+- Cloud-native deployment
+
+---
+
+## 🧩 Challenges Encountered & How They Were Solved
+
+This project involved several real-world engineering challenges, particularly around environment setup, dependency management, and cloud deployment.
+
+---
+
+### 1. Docker Desktop Stuck on “Starting”
+
+**Problem:**
+- Docker Desktop failed to start on Windows
+- Root cause was virtualization / WSL misconfiguration
+
+**Solution:**
+- Switched to **Ubuntu (dual boot)** for a native Linux environment
+- Continued development without relying on Docker Desktop
+
+**Outcome:** Stable and predictable development environment
+
+---
+
+### 2. AWS Lambda Import Error (Critical)
+
+**Problem:**
+-No module named 'pydantic_core._pydantic_core'
+
+**Root Cause:**
+- Dependencies were built on Ubuntu
+- AWS Lambda runs on Amazon Linux
+- Compiled libraries (like `pydantic_core`) were incompatible
+
+**Solution:**
+- Rebuilt dependencies targeting a Linux-compatible environment
+- Adjusted packaging process
+- Ensured correct deployment structure
+
+ **Key Insight:**
+> Compiled dependencies must be built in an environment compatible with the target runtime.
+
+---
+
+### 3. Lambda Packaging & Structure Issues
+
+**Problem:**
+- Incorrect zip structure
+- Lambda could not locate modules or handler
+
+**Solution:**
+- Ensured all dependencies and files were at the root level of the zip
+- Verified handler configuration (`lambda_handler.handler`)
+
+ **Outcome:** Successful Lambda initialization
+
+---
+
+### 4. CloudFront Integration Challenges
+
+**Problem:**
+- Attempted to use **CloudFront** to serve the application
+- Encountered issues integrating dynamic API responses with static hosting
+- Complexity in routing requests between S3 (frontend) and API Gateway (backend)
+- Additional configuration overhead (origins, behaviors, caching rules)
+
+**Why it failed (practically):**
+- CloudFront is optimized for **static content delivery**
+- This project required **dynamic backend communication (chat API)**
+- Misalignment between use-case and tool added unnecessary complexity
+
+---
+
+### 5. Architectural Pivot → App Runner
+
+**Decision:**
+Instead of forcing a CloudFront-based architecture, the backend was moved to **AWS App Runner**
+
+**Why App Runner:**
+- Designed for running **containerized web services**
+- Handles HTTP traffic natively
+- Eliminates API Gateway + Lambda complexity
+- Simplifies deployment and debugging
+
+**Solution:**
+- Containerized the FastAPI application
+- Deployed using App Runner
+- Configured proper start command (`uvicorn`)
+
+**Outcome:**
+- Faster deployment
+- Fewer moving parts
+- Easier debugging and iteration
+
+---
+
+### 6. Runtime Startup Failures (App Runner)
+
+**Problem:**
+- Application built successfully but failed during deployment
+- No server was listening on expected port
+
+**Solution:**
+- Explicitly defined start command:
+-uvicorn server:app --host 0.0.0.0 --port 8080
+- Ensured correct working directory and module path
+
+ **Outcome:** Application successfully started and served traffic
+
+---
+
+## Deployment Notes
+
+- Backend must expose and listen on port `8080`
+- Start command must explicitly run the ASGI server (uvicorn)
+- Environment consistency is critical for dependency compatibility
+
+---
+
+## 🛠️ Tech Stack
+
+- Python (FastAPI)
+- OpenAI API
+- AWS Lambda
+- API Gateway
+- AWS App Runner
+- Amazon S3
+- Mangum
+- uv (Python package manager)
+
+---
+
+## Future Improvements
+
+- Add CloudFront CDN (properly configured)
+- Improve UI/UX
+- Add authentication
+- Introduce vector database for memory
+- Fully containerize and standardize deployment with Docker
+
+---
+
+## Key Lessons Learned
+
+- Environment consistency is critical
+- Serverless deployments require strict packaging rules
+- Compiled dependencies can break across systems
+- Choosing the right AWS service matters more than forcing one
+- Simpler architectures are often more reliable
+
+---
+
+## 👩🏽‍💻 Author
+
+**Mwende Mugao**
+
+---
+
+## ⭐ Final Note
+
+This project demonstrates:
+- End-to-end AI system development
+- Cloud deployment and architecture decisions
+- Real-world debugging and problem-solving