Muhammad Gohar Javed
Resume
🔍Looking for full-time opportunities in Machine Learning...
I am a Masters student at the University of Alberta working with Dr. Li Cheng and Dr. Xingyu Li at the Vison and Learning Lab. My current work focuses on AIGC for controlling digital human characters and their interactions using Generative AI.
I have two years of professional experience in various Machine Learning roles at AltaML, Hotpot.AI and Teradata. I have worked on various impactful projects for diverse clients spanning the entire Machine Learning Process Lifecycle (MLPL) in Computer Vision and Time Series Analysis. Before that, I recieved my Bachelors degree from National University of Sciences and Technology (NUST), where I conducted research on Neural Network Acceleration for Embedded Hardware.
🔥Recent Paper Release on Arxiv:
InterMask: 3D Human Interaction Generation via Collaborative Masked Modelling
Publications and Projects
-
Applicability of machine learning techniques in predicting wheat yield based on
remote sensing and climate data in Pakistan, South Asia
S Arshad, JH Kazmi, M Gohar Javed, S Mohammed
European Journal of Agronomy, 2023
Paper -
QuantYOLO: A High-Throughput and Power-Efficient Object Detection Network
M Gohar Javed*, M Raza*, MM Ghaffar, C Weis, N Wehn, M Shahzad, F Shafait - *Equal Contribution
Proceedings of the Digital Image Computing: Technqiues and Applications (DICTA), 2021 - Spotlight
Paper
Education
University of Alberta
Master of Science in Electrical and Computer Engineering
Specialization: Signal and Image Processing
Thesis: Controllable Generation of 3D Physical Human Interactions
Advised by: Dr. Li Cheng and Dr. Xingyu Li
CGPA: 4.00
National University of Sciences and Technology (NUST), Pakistan
Bachelor of Electrical Engineering
Thesis: Autonomous Landing of Fixed Wing UAV using Computer Vision Techniques
Advised by: Dr. Faisal Shafait
CGPA: 3.84
Experience
Alberta Machine Intelligence Institute (Amii), Canada
Team Lead: Sahir
- Advised 4 startups, across different industries, on scaling their ML capabilities through coaching sessions with CTOs and ML Leads, as a student consultant for the Level Up program.
- Helped them translate business requirements into ML solutions, brainstorm ideas, and solve challenges using the latest academic research and industry trends.
AltaML, Canada
Manager: Joan Vlasschaert
- Showcased the potential of Machine Learning to reduce unplanned maintenance costs and minimize operational downtime for the vehicle fleet of City of Calgary.
- Developed a Fleet Maintenance Prediction Proof-of-Concept (PoC) using predictive time series analysis on maintenance logs, vehicle sensor data, weather, and road conditions.
Faculty of Engineering, University of Alberta, Canada
Supervised by: Dr. Li Cheng and Dr. Xingyu Li
- Working at the Vision and Learning Lab on digital character control using novel Generative AI methods for 3d human motion and interaction synthesis.
- Published one paper at CVPR 2024 (MoMask) and one recent preprint on Arxiv (InterMask).
Hotpot AI, USA
- Led enhancements for the Object Eraser, Background Removal, and Style Transfer tools, involving full ML process lifecycle to achieve state-of-the-art performance and efficient deployment.
- Enhanced performance quality by 3x for Object Eraser, 5x for Background Removal and 4x for Style Transfer, advancing their competitive position in the market, resulting in increased usage and multiple B2B contracts.
Teradata Global Consulting, Pakistan
Manager: Dr. Salman Shahid
National Oilwell Varco (NOV), USA:
- Enhanced Bitbox, an automated PDC drill-bit inspection solution, by redesigning the cutter detection and grading algorithms to improve its accuracy and reliability.
- Applied advanced 3D segmentation techniques, unsupervised clustering, and geometric data processing to accurately identify and isolate individual drill bit cutters within 3D point cloud data.
- Increased algorithm accuracy from 40\% to 87\%, achieving precise segmentation and grading of drill-bit components, which enabled more effective damage assessment and streamlined inspection processes, and transitioned the project from Proof of Concept (PoC) to Minimum Viable Product (MVP).
- Developed Retail Vision as-a-Service (RVaaS), an intelligent computer vision service to analyze in-store customer behavior by extracting hidden insights from cctv video streams.
- Trained and deployed object detection and tracking models to identify patterns of customer activity for actionable insights.
University of Tokyo, Japan
Supervised by: Dr. Takeshi Morita
- Developed a circuit design prototype to harvest green energy from vibrations using Piezoelectricity.
Selected Awards & Achievements
- CVPR 2023 Student Grant - Received a grant to attend the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition covering conference registration, travel, and accommodation expenses.
- UAlberta Graduate Research Fellowship - Secured a fully funded Masters in Research position at the University of Alberta, Faculty of Engineering. CAD 19,500 per year award.
- Teradata Spot Award - For exceeding expectations in the completion and delivery of Vision Analytics Proof-of-Concept for Resona Bank Japan.
- NUST Dean's Merit Scholarship - Scholarship award for achieving academic excellence in all 8 semesters.
- International Physics Olympiad Switzerland (IPhO) 2016 - Selected among the top 5 students to represent Pakistan after 3 extensive examination camps.
- National Light Olympiad 2016 - Bronze medal for securing 3rd position out of 20 selected participants from all over the country. The Olympiad consisted of theoretical and experimental tests based on Light and Optics.
Talks and Presentations
-
Cohere for AI Community Talks - MoMask (CVPR 2024):