Job Market Intelligence Platform
Full-Stack AI Platform · FastAPI · PostgreSQL · LLM/RAG · Next.js
Backend-first AI platform that ingests, normalizes, and analyzes job-market data to produce
explainable recommendations and market intelligence through async pipelines and a React frontend.
- FastAPI backend with async ingestion and enrichment pipelines
- PostgreSQL persistence with normalized schemas and historical snapshots
- RAG-style enrichment and explainable recommendation workflows
- React / Next.js frontend consuming backend APIs
Multimodal Visual Inspection & Explanation API
FastAPI · VLMs · Multimodal AI · Document AI
Production-style multimodal AI service that analyzes images and documents and returns
structured results with grounded explanations, confidence signals, and recommended next steps.
- Image + prompt → multimodal reasoning → structured output
- Clear separation between perception, reasoning, and interpretation
- Confidence-aware outputs and failure-aware response handling
- API-first backend architecture
Azure Serverless Invoice Extraction API
Azure Functions · Document Intelligence · CI/CD
Serverless API that accepts invoice documents and converts them into structured JSON using
Azure Document Intelligence, with CI/CD and monitoring.
- HTTP-triggered Azure Function with dedicated extraction service layer
- Normalized Azure Document Intelligence output into clean JSON
- GitHub Actions pipeline with testing and deployment checks
Azure Customer Segmentation Pipeline
Azure ML · Batch Endpoints · K-Means · MLOps
End-to-end ML pipeline for customer segmentation using feature engineering, K-Means clustering,
Azure ML Batch Endpoints, and reproducible deployment workflows.
- Feature engineering for recency, frequency, monetary value, and customer behavior
- Model training and registration in Azure ML
- YAML-based batch endpoint and deployment definitions
RAG Microservice with FastAPI & ChromaDB
FastAPI · ChromaDB · SentenceTransformers · Groq
Retrieval-Augmented Generation microservice that ingests documents, indexes embeddings,
performs vector search, and returns grounded answers through a FastAPI `/ask` endpoint.
- Document ingestion pipeline for chunking, embedding, and vector storage
- Online retrieval pipeline with grounded answer generation
- Dockerized service with environment-based LLM configuration