AI
-
Jumped into AI on
OpenAI DevDay 2023.
-
As a product manager, led the project NOVA (RAG over Retail Stores database) built for
INNOVENT Technologies, Demoed at GITEX, GLOBAL
-
Currently building projects at
Headstarter
across AI/ML and full stack web apps. Collaborating in
teams to deploy custom neural networks, Voice AI agents,
Generative UI components, multimodal RAG pipelines, and
recommendation systems
-
Here is my indepth article on
The Foundations and Evolution of Generative AI: A Holistic Overview of Models,
Techniques, and Applications, which can help you digest a lot of AI Concepts
Top AI Projects, I Built
-
AI Search Engine:
Developed using Next.js and TypeScript, enabled web scraping with Puppeteer and Cheerio, and implemented an LLM-based chat interface with source citations. Optimized performance with Redis-based rate limiting and scalable API handling.
-
Pentagram: Realtime Image Diffusion:
Built a web app for generating images from text prompts using serverless GPUs, React, and TypeScript. Achieved low-latency performance (<2 seconds) for concurrent requests with scalable GPU resource management.
-
Stock Analysis and Search Engine:
Designed a system for finding stocks based on natural language queries. Integrated a scalable embeddings pipeline with Pinecone, a FastAPI for text embeddings, and filtering by key metrics like market cap and volume.
-
LLM Evaluation Platform:
Developed a web app to compare LLM responses side-by-side with metrics like accuracy and response time. Built an analytics dashboard for visualizing and comparing experiment results.
-
Brain Tumor Classification:
Classified 1000 MRI scans into three brain disease categories using neural networks. Delivered multimodal reports for neurosurgeons with under 200ms latency after training.
Interesting AI Talks/Videos, You must Watch
Papers
- Low-Resource End-to-End Urdu TTS using Tacotron2, WaveGlow, and Transfer Learning: Link to Download
