Muhammad Shoaib Ahmed Siddiqui, Developer in Cambridge, United Kingdom
Muhammad is available for hire
Hire Muhammad

Muhammad Shoaib Ahmed Siddiqui

Verified Expert  in Engineering

Machine Learning Developer

Location
Cambridge, United Kingdom
Toptal Member Since
April 6, 2020

Shoaib is part of the machine learning group at the University of Cambridge as a PhD student. Shoaib has worked on various machine learning (ML) algorithms. Most of these ideas have been published in leading journals and conferences. With his novel research background, he is extremely up-to-date with the recent papers and research going on around the world in ML; this demonstrates Shoaib's capability to integrate cutting-edge research into production.

Artificial Intelligence (AI)Machine LearningDeep LearningLinuxJavaC++PythonNumPyPyTorchNatural Language Processing (NLP)Computer VisionTime SeriesCJavaScriptOpenCV

Availability

Full-time

Preferred Environment

NVIDIA CUDA, TensorFlow, PyTorch, Linux

The most amazing...

...algorithm that I worked on is regarding data cleansing.

Work Experience

Researcher

2017 - 2021
German Research Center for Artificial Intelligence (DFKI)
  • Co-supervised bachelor's and master's projects and thesis.
  • Created deep learning systems for automated fault detection in the industry (mainly anomaly detection).
  • Developed table detection and structure recognition systems (mostly published as research papers).
  • Demystified deep learning models for time-series analysis through visualization.
  • Created adversarial examples and defenses for time-series and visual modalities.
Technologies: Python, NVIDIA CUDA, TensorFlow, PyTorch

Researcher

2019 - 2019
NVIDIA Research
  • Developed a principled approach towards defending against attacks through better understanding the root cause of the vulnerability of deep learning models against adversarial attacks.
  • Employed alternate visual image representations inspired by the 3D human visual system to enhance adversarial robustness.
  • Defined large-scale deep learning benchmarks with efficient data loading pipelines.
  • Created attention models and weakly-supervised object localization.
Technologies: NVIDIA CUDA, PyTorch, Python

Technical Project Lead

2016 - 2017
RheinMain University of Applied Sciences (HS-RM)
  • Served as the technical project lead for FibeVid project (fish biodiversity estimation by low-cost non-destructive video-based sampling).
  • Developed a deep learning-based segmentation system for real-time monocular object detection and tracking.
  • Led fish classification using cross-layer pooling of deep CNN.
Technologies: NVIDIA CUDA, TensorFlow, C++, Python

Research Assistant

2015 - 2016
TUKL-NUST Research \& Development Center (SEECS, NUST)
  • Performed document detection and analysis using OpenCV and Tesseract - ICDAR SmartDoc competition 2015 (7th Position).
  • Provided document detection and classification using convolutional and LSTM networks.
  • Detected and segmented objects using convolutional neural networks.
  • Contributed to fish detection and classification using classical computer vision approaches coupled with deep learning.
Technologies: TensorFlow, C++, Python

Cross-Layer Pooling of CNN Activations

https://github.com/shoaibahmed/CrossLayerPooling
Cross-layer pooling algorithm (Liu et al. 2015) in C++ (using OpenCV and Caffe) and Python (using TensorFlow). The code uses a pretrained ResNet-152 network for its initialization. Refer to the paper for more details on ResNet (He et al. 2015) [https://arxiv.org/abs/1512.03385] and [https://arxiv.org/abs/1411.7466] for details on the cross-pooling method.

FCN in TensorFlow

https://github.com/shoaibahmed/FCN-TensorFlow
One of my first implementations of FCN using TensorFlow. Used the most powerful base model of Inception ResNet v2 at that time. The project received some attention on GitHub at that time when there was a dearth of good FCN implementations using TF.

TorchBrain

https://github.com/shoaibahmed/torchbrain
Implemented a computational model of the natural emergence of pooling layers in the for visual recognition between the retina and the LGN during the development of the human brain. The project was based on a NeurIPS-19 paper on this topic.

Deep SVDD

https://github.com/lukasruff/Deep-SVDD-PyTorch
A deep anomaly detection method inspired by the support vector data description (SVDD) where the feature projection using kernel methods was replaced by a deep network initially trained as an auto-encoder.

Languages

Java, Python, C++, C, C#, JavaScript

Libraries/APIs

NumPy, PyTorch, OpenCV, TensorFlow

Platforms

Linux, NVIDIA CUDA, Amazon Web Services (AWS)

Other

Image Recognition, Artificial Neural Networks (ANN), Artificial Intelligence (AI), Machine Learning, Deep Learning, Image Processing, Natural Language Processing (NLP), Computer Vision Algorithms, Computer Vision, Time Series, GPT, Generative Pre-trained Transformers (GPT), Convex Optimization, Lambda Functions, Diffusion Models, Stable Diffusion, ControlNet, 3D

Tools

Jupyter

Paradigms

Distributed Computing

2022 - 2023

PhD in Information Engineering

University of Cambridge - Cambridge, UK

2017 - 2021

Master of Science Degree in Computer Science (Intelligent Systems)

TU Kaiserslautern - Kaiserslautern, Germany

2012 - 2016

Bachelor of Science Degree in Computer Science

National University of Sciences and Technology - Islamabad, Pakistan

OCTOBER 2016 - PRESENT

Machine Learning

Coursera

APRIL 2016 - PRESENT

Machine Learning: Classification

Coursera

JANUARY 2016 - PRESENT

Machine Learning: Regression

Coursera

DECEMBER 2015 - PRESENT

Introduction to Game Development

Coursera

OCTOBER 2015 - PRESENT

Machine Learning Foundations: A Case Study Approach

Coursera

Collaboration That Works

How to Work with Toptal

Toptal matches you directly with global industry experts from our network in hours—not weeks or months.

1

Share your needs

Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2

Choose your talent

Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3

Start your risk-free talent trial

Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring