Muhammad Raffay Khan

Academic

Verbatim text

Muhammad Raffay Khan
-
+92 349-3662991, m .raffa~khan121@gmail.com
~
G-11/2, Islamabad.
Linkedln: httg_s:L[www.linkedin.com[in[raffaykhan2502l
Education
Bachelor of Science (Artificial Intelligence)
Major:
Machine Learning, Artificial Intelligence, Natural Language Processing, Neural Networks and
Deep Learning, Knowledge Graphs, Digital Image Processing, Design and Analysis of Algorit hms,
Database Systems, Computer Vision, Game Theory and Multiagents, MLOps
Faisalabad Grammar School, Faisalabad, Punjab
~-Level (Physics, Chemistry, Mathematics)
Faisalabad Grammar School, Faisalabad, Punjab
P-Level (Physics, Chemistry, Biology, Mathematics)
'Projects
!final Project: SnapShop [Python, StableDiffusion, HuggingFace, Git, Flask]
~ web app for Al-powered generation of images and video advertisements using clothing items and human models
provided by the user and prompt-guidance for custom poses and environemnts.
lproiects:
Stock Market Prediction using Reinforcement Learning [FinRL, Pyfolio, QuantStats] :
Financial RL(FinRL) Library was used to train and then backtest on data in two phases: during normal times and during
market downturn to avoid alpha decay
Brain Tumour Segmentation [Tensorflow, Sklearn, NumPy, Sklmage, Keras]:
~ Unet based GAN architecture was implemented to segment the Tumour given the Tumour images and the mask.
Further techniques were applied to determine its size for a real-world application.
RASA + GPT-J chatbot [HuggingFace):
~ chatbot using RASA was fine-tuned on an exclusive dataset and the GPT-J was used as a fallback.
Facial Recognition System [PyQt, PyTorch, OpenCV) :
~ Facial Recognition system was developed using siamese neural networks and PyQt was used to develop the frontend
and runtime recognition.
gMLP for image classification [Numpy, torch, PIL):
~ gated Multi-layer preceptron was implemented and used to classify the CIFAR-10 dataset. The results were then
;:ompared with vision Transformers without self attention.
Work Experience
Machine Learning Engineer, Adology.ai, USA(Remote).
Aug 2024 - Present
Involved with a variety of tasks mainly related to videos including but not limited to video
similarity, text-to-video retrieval, video captioning, prompt summarization, text-to-speech,
etc. Currently working on two modules: video ad generation through both text-to-video
models and image-to-video models as well as conventional methods such as GAN's, FOL
models, etc. and Knowledge graph (Neo4j) & Graph RAG for data summarization.
Machine Learning Engineer, The Metaverse Company, UK(Remote).
Apr 2024 - Aug 2024
~ctively Involved in Al team, working on state-of-the-art Generative Al methods including
lmage-to-3D, Efficient 3D meshes, and VTON, as well as their deployment on cloud servers
such as Vast.ai, Runpod.io and Vertex.ai.
'Skills & Tools
Professional Skills
Managing Innovation, Agile, Scrum, Technical Leadership, Communication, Product
Development, Planning and scheduling, Quality management, Resource management,
Communication management, Stakeholder management, Strategic thinking, Innovation,
Decision Making.
iTechnical Skills
Python, TensorFlow, Keras, PyTorch, R, MATLAB, SQL, SPARQL, GraphDB, OWL, ComfyUI, LLM's,
RAG, Llamalndex, REST API, FAST API, Crew.ai, Azure, AWS
'Achievements
~warded twice for semester Academic Performance.
Listed on Dean's List of Honor and Reactor's List of Honor for Academic Performance.
FAST NUCES ISLAMABAD CAMPUS

AI enrichment