Hi, I'm Saad 👋
Deep Learning | NLP | Computer Vision | I love building things and solving real-world problems.
Saad

About

I am a final-year student at FAST NUCES with a strong interest in Deep Learning, NLP, and Computer Vision. My journey includes working on AI research, building intelligent systems, and experimenting with cutting-edge models. I enjoy solving real-world problems through technology and continuously expanding my skill set.

Skills

Python
PyTorch
TensorFlow
Machine Learning
Deep Learning
NLP
Computer Vision
Transformers
Generative AI
Hyperspectral Imaging
Geospatial Data
Recurrent Neural Networks (RNN)
Retrieval-Augmented Generation (RAG)
Fine-Tuning
Prompt Engineering
FastAPI
Web Scraping
Linux
Git
SQL
Databases
Data Science
Data Mining
GeoAI
My Projects

Things I’ve made trying to put my dent in the universe.

I’ve worked on tons of little projects over the years but these are the ones that I’m most proud of. Many of them are open-source, so if you see something that piques your interest, check out the code and contribute if you have ideas for how it can be improved.

Ad-Pilot

Created a marketing automation platform tailored for small businesses to manage ad campaigns across Facebook and Instagram.Ad-Pilot streamlines content creation, automated scheduling, real-time performance insights, and competitor analysis to enhance marketing efficiency.

Next.js
FastAPI
PostgreSQL
FLUX.1 [Schnell]
Llama 3.2
Meta Graph API
Serper API

GRU Based Roman Urdu Ghazal Generation

Created a deep learning-based Roman Urdu poetry generator using GRU neural networks. It generates ghazals in the style of renowned Urdu poets by leveraging datasets scraped from Rekhta and features an interactive Streamlit web interface.

Gated Recurrent Units (GRU)
Natural Language Processing (NLP)
Streamlit
PyTorch
BeautifulSoup
Selenium

Sarcasm Detection Tool for Urdu Text on Social Media

Developed an Urdu sarcasm detection tool that leverages a Gaussian Naive Bayes model trained on Urdu text data. It features a custom preprocessing pipeline for text normalization, stopword filtering, and tokenization to analyze social media content.

Machine Learning
Natural Language Processing (NLP)
TensorFlow
Gaussian Naive Bayes
Python
Streamlit

WhatsWhisper

Created a multi-featured WhatsApp bot that transcribes voice messages into text using advanced models like OpenAI's Whisper and Alibaba's ZipEnhancer. Ideal for noisy environments or when listening isn't possible, while also offering smart task scheduling and leveraging Acoustic Noise Enhancement for improved transcription accuracy.

OpenAI Whisper
Alibaba ZipEnhancer
Venom Bot
FastAPI
Python
Phi 3.5

Text Summarization Using Large Language Models

Developed a state-of-the-art text summarization system by fine-tuning BERT, GPT-2, and Llama to generate both extractive and abstractive summaries, leveraging advanced large language models in a Jupyter Notebook environment.

BERT
GPT-2
Llama
Jupyter Notebook
Pytorch
PEFT
LORA
QLORA

Arabic To English NMT Using Transformers

Developed an efficient Arabic-to-English Neural Machine Translation model from scratch by harnessing state-of-the-art Transformers. Applied innovative techniques to deliver high-quality language translation.

Transformers
Streamlit
Sentence Piece
Pytorch
Publications

Research Contributions & Academic Work

A collection of my research papers and academic contributions in the fields of Deep Learning, Hyperspectral Imaging, and Computer Vision. These publications represent my commitment to advancing the field through rigorous research and innovative solutions.

EnergyFormer: Energy Attention with Fourier Embedding for Hyperspectral Image Classification

Saad Sohail, Muhammad Usama, Usman Ghous, Manuel Mazzara, Salvatore Distefano, Muhammad Ahmad

IEEE Geoscience and Remote Sensing Letters (IEEE GRSL) (2025)

Submitted

This novel approach introduces Energy Attention, a mechanism designed to enhance feature learning by integrating energy spectral density into the attention process. Additionally, it leverages Fourier Embedding to capture spatial-spectral correlations, effectively improving classification performance.Why does this matter? Traditional attention mechanisms often struggle with the high-dimensional complexity of hyperspectral data. EnergyFormer overcomes this challenge by incorporating spectral energy information, making it more robust, efficient, and accurate.

Contact

Get in Touch

Want to chat? Just shoot me a dm with a direct question on my Email and I'll respond as soon as possible.