Mohammed Yasser H

Mohammed Yasser H  new_releases

About

Detail

AI / ML Cloud Engineer | 5.5 years Exp | MS - Liverpool JM University | 30+ Certifications including 2x AWS, 3x GCP | Notable Clients: Apple, Google & Qualcomm.
India
Contact Mohammed regarding: 
work
Full-time jobs
Starting at USD7k/month
Flexible work
Starting at USD68/hour

Timeline

work
Job
school
Education
folder
Project
flag
Award

Résumé

Jobs verified_user 67% verified
  • A
    Sr. AI/ML Cloud Engineer
    Affine analytics
    Nov 2022 - Current (3 years 7 months)
    • Provisioned as Senior Consultant vendor for Apple client on-site for the R&D team of the iPhone Product Ops. • Led automation of ML prototyping pipelines, reducing development overhead 85%, cutting iPhone manufacturing costs by ~10%. • Built & deployed GenAI RAG chatbot, while reducing query latency by 45% and improving retrieval accuracy 30%. • Integrated models via REST APIs on AWS, reducing deployment cycles by 40% and lowering infra costs by 20%.
  • Mediaagility
    Sr. AI & Machine Learning Engineer verified_user Verified experience
    Mediaagility
    Aug 2021 - Oct 2022 (1 year 3 months)
    • Administered the role of AI & ML Engineer, providing cloud solutions & services to clients on GCP, AWS, & Azure. • Successfully delivered $2.5+ million Google Research funded projects for establishment of ML Cloud infra for major clients. • Strategized development of end-to-end ML pipelines, reducing the average costs by 21% & boosting research stats by upto 75%.
  • Veoneer
    AI & Machine Learning Engineer verified_user Verified experience
    Veoneer
    Nov 2019 - Jul 2021 (1 year 9 months)
    • Supervised the role of AI & ML Engineer in the R&D Team of Autonomous Driving & Safety Devices. • Successfully delivered top-quality reports for 6+ projects, while maintaining OTD, ensuring a 100% Client retention rate. • Deployed end-to-end Machine Learning Models into production on V-Net Cloud, improving automation workflow by 60%. Additionally, explored vector similarity techniques for object detection and feature matching, which align with concepts used in vector databases, enhancing the overall effectiveness of our machine learning applications.
Education verified_user 100% verified
  • L
    Master of Science verified_user Verified experience
    Liverpool Jhon Moores University
    Mar 2021 - Jun 2023 (2 years 4 months)
  • Nitte Meenakshi Institute of Technology
    Bachelor of Engineering verified_user Verified experience
    Nitte Meenakshi Institute of Technology
    Aug 2015 - Oct 2019 (4 years 3 months)
  • G
    Google Cloud Professional Machine Learning Engineer Certification verified_user Verified experience
  • G
    Google Cloud Associate Engineer Certification verified_user Verified experience
Projects (professional or personal) verified_user 84% verified
  • Apple
    E2E ML Pipeline Manufacturing Production lines
    Apple
    Nov 2022 - Current (3 years 7 months)
    • Tech used: Python, TensorFlow, PyTorch, FastAPI, Plotly, FastML, Datalab, DataRobot, AWS, Serin, Tableau, HTML. • Built E2E ML pipelines for rapid training & deployment with monitoring, to effectively detect the failures in the manufacturing & assembly process, reducing the need for manual inspection and thus bringing down the overall manufacturing cost by ~10%. • Increased the pace of rapid prototyping of ml models, by automating the workflow, saving the 85% modelling overhead. • Optimized AutoEncoder based Anomaly detection models to effectively detect outliers that lead to breaking of production line.
  • O
    Protein Sequence Model Optimization & Cloud Migration verified_user Verified experience
    Ordaos
    Feb 2022 - Oct 2022 (9 months)
    • Tech used: Python, TF, PyTorch, Azure - ML, VMs, Blob Storage, Git, GCP - Vertex AI, GCS, CE, BigQuery, AI Platform, CSR, GCR. • Streamlined the migration of ML pipelines from Azure Infrastructure to Google Cloud Platform. • Restructured the ML code to effectively utilize TPUs' & Nvidia GPUs' setup, reducing 12% training time & achieving upto 18% cost savings. o Optimized the Perceiver IO Model integration to effectively utilize the existing hardware setup. (Multithreading, FP, Map). o Created custom TFRecords data-source with high throughput protocol buffer reducing the ram memory usage to minimal. o Conducted considerable benchmark tests to minimize the overhead for selecting optimal hardware configuration. o Employed Reduction Servers
  • Persistent Systems
    Computer Vision based Document & ID Scanner verified_user Verified experience
    Persistent Systems
    Aug 2021 - Jan 2022 (6 months)
    • Tech used: Python, TensorFlow, OpenCV, OCR, Flask, AWS – Rekognition, Textract, SageMaker, Google Cloud Platform - Vertex AI, Vision API, Document AI, GCS, BigQuery, AppEngine Flexible, Cloud Source Repository, Container Registry. • Built a Computer Vision based ID Document Scanner using Deep Learning to capture, scan, extract & store the information on Cloud Database in real-time. • Prototyped end-to-end pipeline with a WebApp interface using Flask API, which is estimated to save 70% of the capital resources. • Emphasized great value in research & development, by attracting 2+ client project interests, expanding anticipated proposals.
  • Qualcomm
    Advancement of Mono-Vision Camera Systems for Autonomous Driving - (Client: Qualcomm) verified_user Verified experience
    Qualcomm
    Jan 2021 - Jul 2021 (7 months)
    - Accelerated Computer Vision Advancements to improve existing pipelines with applications of Deep Learning using CUDA with Nvidia GPUs. - Improved advancements include Object Detection, Lane Detection, Face Detection & Sign-Board Detection, using advanced custom OpenCV modules. - Resulted in a 60% increase in R&D resources, bringing focus towards achieving autonomy. - Integrated MLOps practices in the development of deep learning models, ensuring efficient model deployment and monitoring. - Utilized tools such as Kubernetes, GCP, and vector databases in related projects, enhancing overall project management and data handling capabilities.
  • General Motor
    Projection of Range for 144Hz Radar 2.0 verified_user Verified experience
    General Motor
    Jun 2020 - Dec 2020 (7 months)
    - Converted raw signal into a useful resource for ML model development, predicting a comprehensive range for a 180-degree signal reach. - Deployed a web-based application into production, significantly improving the overall success rate of experiments from 50% to 90%. - Applied embedding-based signal representations relevant to vector database concepts, enhancing data management and retrieval processes. - Leveraged containerized deployments that align with Kubernetes-based orchestration, ensuring scalability and efficiency in the project environment. - Engaged with startup methodologies to foster innovation and adaptability within the project, utilizing technologies such as Python, TensorFlow, XGBoost, Flask, and HTML for effective web
  • Ford
    Device Failure Prediction System verified_user Verified experience
    Ford
    Nov 2019 - May 2020 (7 months)
    - Recognized and initiated the need for ML Automation in Structural, Thermal, CFD, and EM simulations. - Designed and developed Deep Learning Models to analyze the simulation data, depicting up to 36 features and 4 simulation types. - Achieved an overall F1-Score of 92% on the test sets, which helped to reduce simulation costs and time by 67%. - Although CI/CD, Kubernetes, and vector databases were not directly involved in this project, my broader experience in cloud-based ML projects has equipped me with a comprehensive understanding of these technologies, enhancing my ability to deploy and manage machine learning solutions effectively. Additionally, my experience with Python, TensorFlow, and Scikit-learn, along with web technologies li
Awards verified_user 100% verified
  • Veoneer
    Above & Beyond Honor verified_user Verified experience
    Veoneer
    Jul 2020 - Sep 2020 (3 months)
    - Solely Managed Project with exceptional client satisfaction.