LLM Pipeline v2
Improving RAG chunking strategy with hybrid retrieval
Building & exploring
Muhammad
Asad Khan
|
CS Student @ UBIT '28
Python · AWS · Data Science · AI/ML · Networking
Activity Hub
🎬
Last Watched Film
Checking Letterboxd...
📺
Currently Watching Series/TV Show
Checking Media...
📖
Currently Reading
Checking Library...
Watching Football
Loading scores...
At a Glance
Projects Built
4+
Certifications
6+
Technologies
20+
Build Focus
Solution-oriented Projects
AI Integration & RAG Systems
Cloud Architecture (AWS)
scroll
Who I Am
About Me
I'm a Computer Science student at UBIT, University of Karachi '28, driven to explore every corner of CS before settling on a specialty. So far I've covered Cloud Computing, Networking, Web Dev, Data Science, Version Control, Project Management, Prompt Engineering, and AI — with lots more ahead.
I learn by building. Every project is a live experiment — from AWS VPC architecture to RAG-powered LLM pipelines. My philosophy: build broken things, understand why they broke, then spend ages unbreaking them.
Currently Exploring
active
DSA in C++
Deep diving into beginner data structures and algorithmic complexity
Queuing Systems (C#)
Implementing single and multi-server queues using real-world scenarios
Work
Projects
Asaaniyat
An agentic AI mobile orchestrator organizing Pakistan's informal economy. Features a custom 8-agent pipeline automating intent parsing from Roman Urdu voice inputs, provider discovery, and dynamic pricing.
Mogscope
Facial analytics platform using face-api.js for 68-point landmark detection. Combines ML insights with LLM-generated satirical analysis and voice synthesis.
LegalEaseAI
Upload contracts for rapid AI risk-rating. Provides plain Urdu explanations, a RAG-powered document chatbot, and downloadable PDF reports. Powered by FastAPI and LangChain.
PollPulse
Cloud-Deployed Capstone
Real-time polling platform built on AWS. Features a custom VPC with EC2 and RDS MySQL isolation. Delivers live data visualization via Chart.js and secure Node.js authentication.
Career
Experience
Under active construction — laying down the first bricks.
Arsenal
Tech Stack
Digital workspace — categorized by core competency
Languages
Core systems & programmable logic
Web Development
Universal frontend & runtime technologies
Cloud & Infrastructure
Infrastructure, Linux Systems & Cloud services
AI / Data
Machine Learning models & Data analysis tools
Backend & Tools
Hybrid APIs & development environments
Background
Education
In Progress (25%)
B.S Computer Science
Department of Computer Science UBIT, University of Karachi
Intermediate (Pre-Engineering)
Govt. Degree Science/Commerce College, Gulshan Block 7
Matric (SSC)
Karachi Public School
Connect