• 🎓 CS undergraduate at SZABIST, Karachi — Batch 2028
• 🏗️ Building backend, cloud, and systems foundations from first principles—understanding how technologies work beneath the abstraction.
• ☁️ Hands-on with AWS, Linux, NGINX, PostgreSQL — deployed production-ready applications on cloud infrastructure.
• 🎯 Long-term focus on Cloud & DevOps Engineering — infrastructure, systems, and reliability at scale
• ♟️ Chess player — drawn to strategy, problem-solving, and long-term thinking.

• Built an AI-powered semantic search engine that enables users to discover movies through natural language descriptions, emotions, and storylines instead of relying on exact keywords or titles.
• Implemented vector search using Google Gemini Embeddings and PostgreSQL pgvector, storing 3072-dimensional embeddings and retrieving semantically similar movies using cosine similarity, ranked by match score.
• Sourced and ingested movie metadata — posters, ratings, cast, and genres — from the TMDB API, integrating rate-limited external data into the embedding and relational pipeline.
• Engineered a Retrieval-Augmented Generation (RAG) pipeline that combines semantic retrieval with Google Gemini to generate contextual explanations for recommendations, supporting English and Roman Urdu movie queries with cross-lingual semantic matching.
• Designed and normalized a PostgreSQL database with 16 interconnected tables, modeling users, movies, genres, cast, ratings, reviews, favorites, watchlists, and search analytics while maintaining referential integrity.
• Containerized the full stack into three independent Docker services (Nginx frontend, Node.js/Express backend, PostgreSQL+pgvector database) orchestrated with Docker Compose, and deployed to a DigitalOcean VM for a reproducible, one-command setup.

• Built a Meeting Intelligence platform that captures live lectures and meetings via microphone, browser tab audio, or file upload — and transforms them into summaries, action - items, key decisions, quiz questions, and follow-up email drafts automatically
• Integrated Google Cloud Speech-to-Text with speaker diarization for multi-speaker transcription and FFmpeg for audio extraction across MP3, MP4, WAV and M4A formats
• Connected Google Gemini 2.0 Flash to analyze full transcripts and return structured JSON output covering all meeting intelligence features
• Built a production-grade Express REST API with four endpoints and automatic retry logic for rate limiting, deployed frontend and backend independently on Google Cloud Run via containerized Docker builds
• Gained hands-on experience with real-time audio processing, AI pipeline architecture, Cloud Run containerization, and full-stack production deployment on Google Cloud

• Built a Linux Bash-inspired interactive shell from scratch in x86 Assembly Language — simulating how real operating systems handle user commands under the hood
• Implemented 8 core shell commands — ls, cd, mkdir, echo, clear, help, exit — alongside file handling operations including create, write, and rm, closely mirroring real Unix shell behavior
• Designed the command parsing and input processing logic manually using string comparison and memory buffers — without any standard libraries or high-level abstractions
• Running on DOSBox, the shell handles real filesystem operations — creating, deleting, and navigating directories and files — through direct DOS system calls
• Strengthened foundational understanding of how operating systems work, how commands are processed at the hardware level, and how memory and CPU registers drive program execution

• Engineered an interactive portfolio with an integrated AI chatbot assistant, enabling real-time conversations about projects, skills, and technical background
• Designed a custom knowledge dataset (knowledge.json) to provide contextual, personalized responses using LLM APIs and optimized prompt handling
• Built full-stack architecture with HTML, CSS, Vanilla JavaScript frontend and Node.js + Express backend handling API routing, request processing, and AI integration
• Deployed production-grade application on a Linux cloud VM, configured NGINX reverse proxy for secure routing, caching, and performance optimization
• Secured deployment using custom domain + SSL/TLS (Let’s Encrypt), enabling HTTPS and production-level security standards
• Gained hands-on experience with cloud infrastructure, server configuration, reverse proxies, environment management, and production deployment workflows

• Built an automated fire detection and alert system integrating IoT flame sensors with a Linux-based monitoring environment
• Implemented event-driven Bash scripts to continuously monitor sensor signals and trigger automated alert workflows
• Configured SMTP-based email notifications to instantly inform relevant authorities when fire was detected
• Designed a lightweight monitoring pipeline using Linux services, logging mechanisms, and automated alert triggers
• Gained hands-on experience in hardware–software integration, Linux automation, event-driven system design, and real-time alerting workflows

• Built a full-featured voting system with admin and voter roles, secure vote casting, and real-time vote counting
• Integrated PostgreSQL database for persistent, structured storage with proper schema design and validation
• Migrated local database to a cloud-hosted PostgreSQL instance on DigitalOcean, enabling remote access and live deployment
• Applied OOP principles, database connectivity (JDBC), and production-style cloud database management

• Ride-sharing platform simulation with driver & passenger management
• Features: ride booking, fare calculation, ride history, admin controls
• Highlights: OOP, data structures, file handling, algorithmic logic

🎯 Currently Learning

• Core backend internals — HTTP lifecycle, TCP/IP, socket communication, request/response model
• AWS core services, IAM, compute, storage, and networking fundamentals

• 💼 GitHub: https://github.com/Pritam-Kumar-911
• 📍 Karachi, Pakistan