---

Saathi – Distributed Coding Platform

Saathi is a production-grade, microservice-oriented platform engineered to facilitate competitive programming (CP) and technical interview preparation. The system is architected to handle high-concurrency code execution while simultaneously providing deep, insight-heavy analytics through a decoupled, asynchronous processing model.

Project Overview

Saathi addresses the inherent challenges of online judge systems: balancing the intensive CPU demands of code compilation and execution with the data-intensive nature of performance analytics. By segregating the system into two distinct operational pipelines—the Dual Practice Pipeline and the Analytics & Review Pipeline—Saathi ensures that compute-heavy workloads do not interfere with the responsiveness of the user-facing analytical services.

Core Design Philosophy

The architecture is governed by three primary principles:

• Isolation of Concerns: Code execution is strictly sandboxed and isolated from the primary application logic to ensure security and resource stability.
• Workload-Specific Optimization: The system utilizes different communication protocols (gRPC vs. REST) and persistence layers (SQL vs. NoSQL) based on the specific needs of the workload.
• Asynchronous Resilience: Long-running execution tasks are handled via message queues to prevent blocking the request-response cycle of the API Gateway.

---

High-Level Architecture Explanation

The Saathi architecture utilizes a centralized API Gateway as the single ingress point for all client traffic. Upon receiving a request, the gateway routes the traffic into one of two specialized pipelines:

1. Dual Practice Pipeline: Manages the lifecycle of code submission, evaluation against test cases, and result reporting.
2. Analytics & Review Pipeline: Manages persistent storage of submission history, AI-powered code reviews, and user growth metrics.

---

✨ Key Features

• 🤖 AI Concept Tagging: Automatically identifies if mistakes are related to specific domains such as Promises, CSS Grid, Logic, or Security.
• 📝 Structured Feedback: Delivers mentor-style explanations, optimized solutions, and curated learning links based on execution results.
• 📊 Personalized Analytics: A specialized dashboard that tracks progress and identifies "Top Struggle Areas" using historical submission data.
• 🔐 Mistake Vault: Every review is persisted to a private history, enabling longitudinal tracking of user improvement.

---

🛠️ Tech Stack

• AI Engine: Gemini 2.0 Flash (Google Generative AI)
• Backend: Node.js & Express (ES Modules)
• Database: MongoDB & Mongoose (Read-optimized for Analytics)
• Execution Environment: Docker Containerization
• Authentication: JWT (JSON Web Tokens) & Bcrypt

---

Workload Breakdown

Workload 1: Code Execution Service (Dual Practice Pipeline)

• Async Task Queue: Decouples submission from evaluation, protecting the system from traffic spikes.
• Sandboxed Docker Execution: Each submission runs in a hardened, ephemeral container with strict CPU/Memory quotas and no egress network access.
• gRPC Return Path: Low-latency, strongly typed communication for immediate execution status delivery to the Gateway.

Workload 2: Code Review & Analytics (Insight Pipeline)

• MongoDB Persistence: Chosen for its schema flexibility to store varied AI-generated feedback and metadata.
• Grok/Gemini Review Service: Performs static and dynamic analysis to provide time complexity and logic optimization insights.
• Centralized Response Handling: The API Gateway aggregates data from both workloads to provide a unified user response.

---

🚀 Getting Started

1. Prerequisites

• Node.js (v18+)
• MongoDB (Local or Atlas)
• Gemini API Key (Get it here)

2. Installation

# Clone the repository
git clone https://github.com/yourusername/saathi-backend.git

# Navigate into the project
cd saathi-backend

# Install dependencies
npm install

3. Environment Setup

Create a .env file in the root directory:

PORT=3000
MONGODB_URI=your_mongodb_uri
GEMINI_API_KEY=your_api_key
JWT_SECRET=your_secure_secret

4. Execution

# Development mode (with Nodemon)
npm run dev

# Production mode
npm start

---

Scalability, Security, and Extensibility

• Scalability: The Code Execution Service supports independent horizontal scaling. During peak contest load, the worker pool can be expanded without impacting the Analytics API.
• Security: Runtime security is enforced via restricted system calls (Seccomp) and resource limits within the Docker sandbox.
• Extensibility: The modular gRPC interface allows for the addition of new evaluation engines (e.g., specialized ML model evaluators) with minimal changes to the core gateway.

---