Experience

Contilio

• United Kingdom
• IT Services and IT Consulting
• 1 - 100 Employee

• Computer Vision / Software Engineer

  • Jul 2021 - Present

  At Contilio, I have had the opportunity the work on real-time progress reporting platform using LiDAR (Point Cloud Data) and BIM (Building Information Modeling). - Technical Lead for multiple customer projects, responsible for streamlining processes and improving infrastructure for releases and automation. - Successfully reduced the time for delivering customer analysis from 72 hours to 24 hours by optimizing pipeline stability and performance. - Product Owner of Augmentation Pipeline, responsible for developing and generating PC data from BIM files for development tasks and customer demo purposes. - Improved labeling process as a Labeling Lead, resulting in a 30% reduction in time spent on labeling tasks. - Contributed to the improvement of classification, resulting in more accurate inference saving upto 42% time in Quality Control and a 70% reduction in processing time. - Strong expertise in process optimization and infrastructure development. - Demonstrated ability to lead and manage multiple projects simultaneously. - Proficient in project management tools and techniques, including agile methodologies. Excellent communication and collaboration skills, able to work effectively with cross-functional teams. Show less



Automotive Artificial Intelligence (AAI) GmbH

• Germany
• Motor Vehicle Manufacturing
• 1 - 100 Employee

• AI Research Scientist

  • Nov 2020 - Jul 2021

  As a Research Scientist at Automotive AI, my job was to develop state-of-the-art technologies for Scenario Cloning and Extraction. The aim was to reduce the gap of real driving experiments and virtual tests, leading to the efficient development of automated driving functions. - Designed and Implemented real-time handcrafted methods to perform Object Detection, Classification and Tracking for point cloud data (LIDAR Data). - Successfully developed the method to predict the Split Point of a vehicle. Split point reduces the error in estimating the orientation, distance and speed of a Vehicle. Show less



CloudSight Inc.

• United States
• Technology, Information and Internet
• 1 - 100 Employee

• Data Scientist

  • Jul 2017 - Nov 2020

  In this role, my responsibility was to lead the development of Computer Vision and NLP projects. At CloudSight I have worked on Scene Understanding, Image Captioning, Image and Text Recognition, Action Recognition in Videos and Language Modeling. - Significantly improved Caption Generation / Scene Understanding pipeline and successfully deployed the models for CamFind, TapTapSee and CloudSightAPI. - Created diverse image descriptions on more than 279K test images to improve accuracy of baseline models. The newfound precision of these models helped us save approximately 250K dollars every month from manual labelling costs. - Reduced memory footprint of Captioning Models up to 55%, allowing on-device inference and a significant decrease in billing costs at Cloud Platforms. - Proposed and incorporated CFRM to improve Landmark Recognition for mobile OEM. - Lead the development of an end-to-end pipeline for Action Recognition in Videos which resulted in state-of-the-art performance on publicly available datasets. - Responsible for supervising and defining architectures for Machine Learning projects. Show less



LIG - Grenoble Informatics Laboratory

• Research Services
• 1 - 100 Employee

• Computer Vision Intern

  • Feb 2018 - Jun 2018

  - Identified issues in state-of-the-art Action Recognition methods and proposed Consensus Pooling Network to overcome these limitations. - Our proposed network achieved state-of-the-art performance on publicly available datasets such as UCF101, HMDB51 and Kinetics. - Identified issues in state-of-the-art Action Recognition methods and proposed Consensus Pooling Network to overcome these limitations. - Our proposed network achieved state-of-the-art performance on publicly available datasets such as UCF101, HMDB51 and Kinetics.



LIG - Grenoble Informatics Laboratory

• Research Services
• 1 - 100 Employee

• • Jun 2017 - Sep 2017

  - Extracted music genres from tweets using LDA to develop Music Recommendation System.- Proposed and developed Deep Learning networks to predict/classify Hashtags of a tweet.

• • Feb 2017 - Jun 2017

  - Performed fine-grained Sentiment Analysis of Tweets using Deep Learning.- Successfully implemented and reproduced state-of-the-art experiment results using Recursive Neural Networks and Convolutional Neural Networks (CNN) for Sentiment Analysis.



Recognition Vision and Learning Lab

• Islāmābād, Pakistan

• Research Intern

  • Aug 2015 - Aug 2016

  - During my 1-year tenure, focused on learning and implementing Deep Learning methods to perform Object Detection and Recognition. - Proposed and implemented an end-to-end pipeline for Text Spotting (Text Detection and Text Recognition) from natural scene images. - During my 1-year tenure, focused on learning and implementing Deep Learning methods to perform Object Detection and Recognition. - Proposed and implemented an end-to-end pipeline for Text Spotting (Text Detection and Text Recognition) from natural scene images.



National University of Computer and Emerging Sciences

• Pakistan
• Higher Education
• 700 & Above Employee

• Teaching Assistant

  • Jan 2014 - Jun 2014

  - In this role, my core responsibility was to Evaluate Assignments for courses; Data Structures, Computer Organisation and Assembly Language. - Arranged Tutorial sessions for students to discuss solution of Assignments and Exams. - In this role, my core responsibility was to Evaluate Assignments for courses; Data Structures, Computer Organisation and Assembly Language. - Arranged Tutorial sessions for students to discuss solution of Assignments and Exams.



National University of Computer and Emerging Sciences

• Pakistan
• Higher Education
• 700 & Above Employee

• Instructor

  • Jun 2013 - Aug 2013

  - My responsibilities were to deliver lectures, prepare lab tasks and evaluate workshop project. - Delivered Lectures on both frontend and backend web development technologies like PHP, HTML5, CSS, JavaScript, Bootstrap and SQL. - My responsibilities were to deliver lectures, prepare lab tasks and evaluate workshop project. - Delivered Lectures on both frontend and backend web development technologies like PHP, HTML5, CSS, JavaScript, Bootstrap and SQL.