From 381d037cb3fd436ff2c00932218f701e0b88a2f2 Mon Sep 17 00:00:00 2001 From: Mugao-joy Date: Mon, 13 Apr 2026 18:45:52 +0300 Subject: [PATCH 1/3] geneinsight studio --- .../mugaoprodweek1/GeneInsight.md | 113 ++++++++++++++++++ 1 file changed, 113 insertions(+) create mode 100644 community_contributions/mugaoprodweek1/GeneInsight.md diff --git a/community_contributions/mugaoprodweek1/GeneInsight.md b/community_contributions/mugaoprodweek1/GeneInsight.md new file mode 100644 index 00000000..872dd919 --- /dev/null +++ b/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** +πŸ‘‰ https://gene-insight-studio-4xh4.vercel.app/ + +
+ +![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** β€” clear biological interpretation of findings. + * **Hypotheses & Next Experiments** β€” actionable scientific next steps. + * **Regulatory / Stakeholder Summary** β€” 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 β†’ Structured insight β†’ 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 + +--- + +## πŸ— 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 + +--- + +## βš™οΈ 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. + +--- From df482a229014c903ce4be2f97eeb5b6efaeeafcd Mon Sep 17 00:00:00 2001 From: Mugao-joy Date: Thu, 30 Apr 2026 11:22:06 +0300 Subject: [PATCH 2/3] week2 digital twin --- community_contributions/mugaoweek2prod.md | 220 ++++++++++++++++++++++ 1 file changed, 220 insertions(+) create mode 100644 community_contributions/mugaoweek2prod.md diff --git a/community_contributions/mugaoweek2prod.md b/community_contributions/mugaoweek2prod.md new file mode 100644 index 00000000..3c4d5cae --- /dev/null +++ b/community_contributions/mugaoweek2prod.md @@ -0,0 +1,220 @@ +# Digital Twin β€” 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. + +--- + +## πŸ”— 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** β†’ backend execution (initial approach) +- **API Gateway** β†’ HTTP routing +- **S3** β†’ frontend hosting +- **App Runner** β†’ final backend deployment + +--- + +## System Flow +User β†’ Frontend (S3) β†’ API Gateway / App Runner β†’ Backend β†’ AI β†’ 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 β€œStarting” + +**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 β†’ 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 + +--- + +## πŸ› οΈ 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 + +--- + +## πŸ‘©πŸ½β€πŸ’» 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 \ No newline at end of file From 82f4008f41c003ca43f18a4f03756fcb6e229d53 Mon Sep 17 00:00:00 2001 From: Mugao-joy Date: Thu, 30 Apr 2026 13:50:54 +0300 Subject: [PATCH 3/3] gene in insight studio deployed --- community_contributions/mugaoprodweek3alex.md | 24 +++++++++++++++++++ 1 file changed, 24 insertions(+) create mode 100644 community_contributions/mugaoprodweek3alex.md diff --git a/community_contributions/mugaoprodweek3alex.md b/community_contributions/mugaoprodweek3alex.md new file mode 100644 index 00000000..a676e2fb --- /dev/null +++ b/community_contributions/mugaoprodweek3alex.md @@ -0,0 +1,24 @@ + +GeneInsight Studio is a biotech-focused SaaS application designed to reduce the cognitive load of interpreting genomic and molecular biology data. +Deployed on google cloud + +## Tech Stack + +### Frontend + +* Next.js (Pages Router) +* React + TypeScript +* Tailwind CSS +* @microsoft/fetch-event-source (SSE client) + +### Backend + +* FastAPI +* Uvicorn +* OpenAI Python SDK +* python-dotenv +* Pydantic + +Deployment +Google cloud :https://geneinsight-studio-git-19933188694.africa-south1.run.app/ +azure cloud : https://gene-insight-container--0mhcvkj.wonderfulgrass-08811794.southafricanorth.azurecontainerapps.io/