Projects
AI Career Discussion Assistant
https://uttamchaturvedi9-carrer-discussion.hf.space/
Overview
I built an AI-powered career discussion assistant that reads a person’s LinkedIn profile (exported as a PDF) along with a supplementary summary text file, and uses that information to have informed, context-aware conversations about the person’s career — their background, skills, and experience.
The core problem this solves is automating the tedious process of manually parsing and summarizing profile documents so that information becomes instantly queryable and conversational, rather than something a recruiter, mentor, or the person themselves has to dig through manually.
How It Works
Accepts a LinkedIn profile PDF export as input
Reads a companion summary text file for additional context
Extracts and processes the document content
Powers a conversational interface that can answer questions and discuss the person’s career based on the ingested data
Tech Stack
Python
Deployed on Hugging Face Spaces
PDF parsing / document ingestion
LLM-based conversational layer
Future Roadmap
This project is designed to be extended into a full RAG (Retrieval-Augmented Generation) pipeline, incorporating:
A vector database for semantic search over larger, multi-document profiles
Chunking and embedding strategies for more accurate retrieval
Support for additional document types beyond PDF/text
Why I Built This
Reading through profile documents to extract relevant career insights is repetitive and time-consuming. This project explores how conversational AI can make that information instantly accessible — and lays the groundwork for a more scalable, retrieval-based architecture as the dataset grows.
Docker Deployment - Backend API- Multi-Stage Build
Implemented a production-ready Docker deployment strategy for a backend API using multi-stage Docker builds to optimize image size and security. The deployment architecture includes:
Multi-Stage Build Process: Separate build and runtime stages reducing final image size by excluding development dependencies and build tools
Secure Dependency Management: Integrated private NuGet feed authentication within the build container for secure package restoration
Automated CI/CD Pipeline: Azure DevOps pipeline automating Docker image builds, versioning, and deployments across multiple environments
Container Registry Management: Dual-registry deployment strategy with automated image tagging and promotion from staging to production
API Documentation: XML documentation files included for Swagger/OpenAPI support in containerized environment
Tech Stack: Docker, .NET 8.0, Azure Container Registry, Azure DevOps, ASP.NET Core
Key Achievement: Streamlined deployment process with automated versioning and multi-environment container orchestration.
Software Bill of Material
Automated SBOM & Security Vulnerability Scanning System — Enterprise .NET Applications
Developed an automated Software Bill of Materials (SBOM) generation and vulnerability scanning solution.
Integrated PowerShell automation within Azure DevOps CI/CD pipelines for seamless execution.
Performed weekly automated scans of all NuGet package dependencies.
Identified and classified vulnerabilities by severity level with comprehensive HTML reports.
Automated archival of reports to Azure Blob Storage for long-term audit tracking.
Eliminated manual security audits, ensuring continuous monitoring and compliance visibility.
Key Technologies: Azure DevOps, PowerShell, .NET, NuGet, Azure Blob Storage
Dating APP – Suitable Match Partner
Dating App is a modern web application designed to help users find and connect with suitable matches. It provides a secure, interactive platform where users can create profiles, upload photos, send messages, and express interest through likes. The project demonstrates the full-stack development of a dynamic, data-driven application with strong emphasis on usability, scalability, and security.
🚀 Highlights:
Built using latest web technologies for modern performance
Entity Framework Code-First ensures easy database evolution
Modular architecture for scalability and maintainability
Implements secure and token-based authentication
Fully functional end-to-end application demonstrating full-stack skills
Key Technologies: C#, Angular, Azure, Sql
GymTracker – Personal Workout Companion
What it is: A simple app to plan workouts, track exercises, and see progress over time.
Who it’s for: Anyone who wants a clean, no‑frills way to log gym sessions and stay consistent.
Key features:
Create and save workouts with common exercises
Log sets, reps, weight, and notes
See recent activity and personal bests
Fast sign‑in and secure data storage
My role: Built the mobile app UI/UX, set up the backend APIs, and connected the two for a smooth experience.
What makes it different: Focuses on speed and simplicity—log a set in seconds, without distractions.
Results: Reliable tracking, clearer progress, and less time fiddling with the phone during workouts.
Availability: Designed for Android and iOS. Demo and screenshots available on request.
Key Technologies: C#, Flutter, Sql