Student Projects — DSWallah AI Engineering Portfolio

🤖 AI Agent

Multi-Tool AI Agent

An AI agent that can search the web, query databases, send emails and interact with external APIs — all through natural language.

PythonLangChainGeminiFastAPIPostgreSQLDocker

┌─────────────┐ │ User Input │ └──────┬──────┘ ▼ ┌─────────────┐ │ LLM Router │ ← Gemini / GPT-4 └──────┬──────┘ ▼ ┌──────────────────────────────┐ │ Tool Executor │ │ ┌──────┐ ┌──────┐ ┌──────┐ │ │ │ Web │ │ SQL │ │Email │ │ │ │Search│ │Query │ │ Send │ │ │ └──────┘ └──────┘ └──────┘ │ └──────────────────────────────┘

📚 RAG System

Legal Document Q&A

Upload any legal contract and ask questions in plain English. Uses hybrid retrieval (vector + keyword) for accurate answers with source citations.

PythonpgvectorOpenAIFastAPIStreamlitRedis

┌──────────────────────────────┐ │ Document Upload │ │ (PDF / DOCX / TXT) │ └──────────────┬───────────────┘ ▼ ┌──────────────────────────────┐ │ Chunking + Embedding │ │ text-embedding-004 / Hugging│ └──────────────┬───────────────┘ ▼ ┌──────────────────────────────┐ │ pgvector + Redis Cache │ │ Hybrid: Semantic + BM25 │ └──────────────┬───────────────┘ ▼ ┌──────────────────────────────┐ │ LLM + Context + Sources │ │ Gemini / GPT-4 │ └──────────────────────────────┘

💬 AI Chatbot

Multi-Model Course Mentor

A RAG-powered chatbot that answers course questions using a knowledge base, with multi-model fallback and lead capture integration.

Node.jsOpenRouterGeminiRAGGoogle Sheets

┌─────────────┐ │ User Query │ └──────┬──────┘ ▼ ┌─────────────┐ │ RAG Search │ ← Knowledge Base └──────┬──────┘ ▼ ┌──────────────────────────────┐ │ Multi-Model Chain │ │ Gemini → Llama → Qwen │ └──────────────┬───────────────┘ ▼ ┌──────────────────────────────┐ │ Response + Lead Capture │ │ WhatsApp / Google Sheets │ └──────────────────────────────┘

📊 ML Pipeline

Customer Churn Prediction

End-to-end ML pipeline: data preprocessing, feature engineering, model training with MLflow tracking, FastAPI serving, and monitoring dashboard.

Pythonscikit-learnMLflowFastAPIGrafanaDocker

┌──────────┐ ┌──────────┐ │ Raw Data │──▶│ Clean + │ │ (MySQL) │ │ Feature │ └──────────┘ │ Engg │ └────┬─────┘ ▼ ┌─────────────────┐ │ Model Training │ │ MLflow Tracker │ └────────┬────────┘ ▼ ┌─────────────────┐ │ FastAPI Server │ │ + Grafana Mon. │ └─────────────────┘

☁️ Cloud Deploy

Production AI App on AWS

Full deployment: Docker containerization, CI/CD with GitHub Actions, AWS ECS/ECS deployment, monitoring with CloudWatch and custom dashboards.

DockerAWSGitHub ActionsFastAPICloudWatchTerraform

┌──────────┐ ┌──────────┐ │ Git Push │──▶│ GitHub │ │ (main) │ │ Actions │ └──────────┘ └────┬─────┘ ▼ ┌─────────────────┐ │ Docker Build + │ │ Push to ECR │ └────────┬────────┘ ▼ ┌─────────────────┐ │ Deploy to ECS │ │ + CloudWatch │ └─────────────────┘

🖼️ Computer Vision

Real-Time Object Detection

Live webcam object detection with YOLOv8, optimized with TensorRT, deployed as a FastAPI service with real-time WebSocket streaming.

PythonYOLOv8TensorRTFastAPIWebSocketDocker

┌──────────┐ │ Webcam / │ │ Video In │ └────┬─────┘ ▼ ┌──────────┐ ┌──────────┐ │ YOLOv8 │──▶│ TensorRT │ │ Detector │ │ Optimized│ └────┬─────┘ └────┬─────┘ ▼ ▼ ┌──────────────────────────┐ │ FastAPI + WebSocket │ │ Real-time streaming │ └──────────────────────────┘

What DSWallah Students Build

AI Agents · RAG Systems · Chatbots · MLOps · Cloud

Ready to Build Projects Like These?